Iowa State University Capstones, Theses and Retrospective Theses and Dissertations Dissertations
1998 Star formation in collisional ring galaxies and the circumnuclear regions of active galactic nuclei Mark Allen Bransford Iowa State University
Follow this and additional works at: https://lib.dr.iastate.edu/rtd Part of the Astrophysics and Astronomy Commons
Recommended Citation Bransford, Mark Allen, "Star formation in collisional ring galaxies and the circumnuclear regions of active galactic nuclei " (1998). Retrospective Theses and Dissertations. 11596. https://lib.dr.iastate.edu/rtd/11596
This Dissertation is brought to you for free and open access by the Iowa State University Capstones, Theses and Dissertations at Iowa State University Digital Repository. It has been accepted for inclusion in Retrospective Theses and Dissertations by an authorized administrator of Iowa State University Digital Repository. For more information, please contact [email protected]. INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.
The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely afreet reproduction.
In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper lefr-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.
Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order. UMI A Bell & Howell Information Company 300 North Zeeb Road, Ann Aibor MI 48106-1346 USA 313/761-4700 800/521-0600
Star formation in collisional ring galaxies and the circumnuciear regions of active galactic nuclei
by
Mark Allen Bransford
A dissertation submitted to the graduate faculty
in partial fulfillment of the requirements for the degree of
DOCTOR OF PHILOSOPHY
Major: Astrophysics
Major Professor: Dr. P. N. Appleton
Iowa State L'niversity
Ames. Iowa
1998
Copyright © Mark .-Vllen Bransford. 199 UMI Microform 9826517 Copyright 1998, by UMI Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. UMI 300 North Zeeb Road Ann Arbor, MI 48103 u Graduate College Iowa State University This is to certify that the Doctoral dissertation of Mark Allen Bransford met the dissertation requirements of Iowa State L'niversity Signature was redacted for privacy. Committee Member Signature was redacted for privacy. Committee Member Signature was redacted for privacy. Committee Member Signature was redacted for privacy. Committee Signature was redacted for privacy. committee Member Signature was redacted for privacy. CommitteeMember Signature was redacted for privacy. K ^ ^ ^Tajor Professor Signature was redacted for privacy. For the Major Program Signature was redacted for privacy. For the Graduate College lit TABLE OF CONTENTS ACKNOWLEDGMENTS vi ABSTRACT x CHAPTER 1 INTRODUCTION I 1.1 A Brief Description of Dissertation Projects and Organization of Introduction . I 1.2 Morphological Classification of Galaxies 2 I..} The Star Formation Properties of Interacting Galaxies 1.4 .-Vctive Gala.xies 0 1.0 Interpretation of Optical Spectra 10 l.o.l Excitation Mechanisms Operating on Xarrow-Emission-Line Gas .... 10 I..5.2 Metal .\bundances in H II Regions i;} 1.6 Dissertation Organization 18 CHAPTER 2 MULTIWAVELENGTH OBSERVATIONS OF COLLISIONAL RING GALAXIES. III. OXYGEN/NITROGEN ABUNDANCES AND STAR-FORMATION PROPERTIES OF RING KNOTS 19 2.1 .-Vbstract 19 2.2 Introduction 20 2.3 The Observations 21 2.4 Determination of the Oxygen and .\itrogen .-Vbimdances and Colors of the Knots 22 2.4.1 E.xtraction of Spectra and Determination of Line Flu.xes 22 2.4.2 Empirical and Ionic abundances of O.xygen and .\itrogen .J? 2.4..} Photometry of Star Forming Knots 2.4.4 Oxygen and .\itrogen .Abundances 415 IV 2.4.5 Evidence for two bursts of star formation in 11 Z\v 28 and .\bell 7fi . . . 51 2.5 Ring Kinematics 52 2.5.1 The Kinematics of the Ring in LT 41 52 2.6 Colors of Star forming Knots and Comparison with .Vumerical Models of Stellar Evolution 55 2.6.1 Description of Isochrone Synthesis .Models 55 2.6.2 Results 58 2.7 Summary and Conclusions 72 2.8 References 73 CHAPTER 3 NEUTRAL HYDROGEN OBSERVATIONS OF ARP 118 . 76 3.1 The .\rp 118 System 76 .{.2 Previous HI and CO observations of Ring Gala.xies 78 VL.\ C-.-Vrray HI Observations of .-Vrp 118 79 3.4 Integrated Properties. HI Distribution, and the Relationship Between the HI and CO emission SI 3.5 HI Kinematics 91 3.6 Discussion 92 3.7 HI Distribution and Kinematics of the Dwarf Companion 95 3.8 Conclusions 100 3.9 References 101 CHAPTER 4 RADIO-LUMINOUS SOUTHERN SEYFERT GALAXIES. 1. RADIO IMAGES AND SELECTED OPTICAL/NEAR-INFRARED SPECTROSCOPY 103 4.1 .Abstract 103 4.2 Introduction 104 4.3 The Sample 105 4.4 The Ob.servations 108 4.4.1 Radio Observations 108 4.4.2 Infrared and Optical Observations 110 4.5 Tiie Radio and FIR Properties 113 V 4.5.1 Radio Morphology lit 4.5.2 Radio Spectral Indices 121 4.5..} FIR/Radio Correlation and FIR Spectral Indices 124 4.6 Optical and .\ear-Infrared Data and its Relation to Radio Structure 129 4.6.1 Optical Spectroscopy of a Subset of the Sample 129 4.6.2 .\ear-Infrared Spectra and Imaging 147 4.6.:J Radio Structure of IC 5063 and its Relation to Published Optical Spectra and Imaging 151 4.7 Conclusions 1-54 4.8 References 157 CHAPTER 5 CONCLUSIONS 160 5.1 Summary of Conclusions 160 5.1.1 Chemical .A-bundances and Star Formation History of Collisional Ring Galaxies 160 5.1.2 .\eutral Hydrogen Observations of .-Vrp 118 162 5.1..3 Circumnuclear Star Formation in .A.G.\s 163 5.2 Ongoing Research and Future Research 164 5.2.1 Ring Gala.Kies 164 5.2.2 Circumnuclear Properties of .AG.\s 165 REFERENCES 167 VI ACKNOWLEDGMENTS "It's not how fast you can go The force goes into the flow If you pick up the beat You can forget about the heat More than just survival More than just a flash More than just a clotted line More than just a dash... In the long run... It's a test of ultimate will The heartbreak climb uphill Got to pick up the pace If you want to stay in the race More than blind ambition .More than simple greed .More than a finish line .Must feed this burning need In the long run... From first to last The peak is never pa^ecl Something always fires the light That gets in your eyes vii One moment'^; high And glory rolls on by Like a streak of lightning That flashes and fades In the summer sky Vbur meters may overload You can rest by the side of the road You can miss a stride But nobody gets a free ride More than high performance More than just a spark More than just the bottom line Or a lucky shot in the dark - In the long run... You can do a lot in a lifetime [f you don't burn out too fast You can make the most of the distance First yoii need endurance First you've got to last..." Marathon, lyrics by .Veil Peart. RL'SH For Lori Christine L'ngerman. "The fawn-eyed girl with sun-browned legs Dances on the edge of his dreams .\nd her voice rings in his ears Like the music of the spheres..." - N'.P. What a life we will have! viii Life can. indeed, be a marathon: it has many significant moments and mile-markers. Ob taining a Ph.D. has been a major goal in my life, and it is a great joy (and relief) that I finish this "heart-break climb uphill". However, there are so many more things I wish to accomplish. This is just the beginning. I have many more challenges ahead, and I look forward to meeting them. Finishing my Ph.D. has as much to with the people in my life that have guided and supported me. as the work done to complete my dissertation. There are those who deserve plenty of credit. [ would like to acknowledge the support and patience of Phil .-Vppieton. my advisor and friend. Phil provided me with a plethora of great projects, and I have had fun (and plenty of angst!) completing them. Curt Struck and Russ Lavery provided me with invaluable advice and guidance, without which my time here would have proved a fruitless endeavor. Curt and Russ have been good friends too: I will miss interacting with them on a day-to-day basis. I am indebted to all three for helping me to e.xpand my horizons as an astronomer, and keeping up my confidence and spirits throughout my time here. I had fun traveling to .Vustralia and Kitt Peak to observe with Tony Marston. I really appreciate the interest Evan Skillnian shows in my work and for asking to be on my committee: it gave me hope and confidence! Lee .-Vnne hiis been very supportive, and a good friend: thanks for alt our great conversations! Ray and Charlene are also good friends, and 1 have had a great time working with them. Hopefully we can continue to work together in the future. I would like to dedicate my thesis to Lori Christine L'ngerman. my fiance and best friend. She has been with me in the final two \-ears leading up to the completion of my dissertation, and has always been unswerving in her support of this endeavor. I have needed to devote many long hours to finishing, and it was a great comfort to know that I had her support and love. In our time together I have founds wells of strength and confidence that I did not know I possessed. I would like to thank my parents for their love and support through the good times and the bad. They were with me at beginning of this quest, and saw me through to the end. They ix have played crucial roles in making me who I am today. .All those trips to planetariums and science museums in .Minnesota burned into me my love of science, and niy in.satiable curiosity about the world that [ carry with me today. To Mike Tammaro. Lars Ewell. and Tony Hill: all the best! In the early days of graduate .school, up until the end. they have been the "gravitational center" of an e.^ctended group which has been the source of many good times and great conversations. They will always be my very best of friends. Finally. I would like to thank Neil Peart for his wondrous and though provoking lyrics. Manna for the soul, truly, while I labored to complete my Ph.D. but also over the years as I learned to live my life. .-Vlso. to RUSH for the hours and hours of music that has comforted me and provided, in a very deef>-rooted sense, a measuring stick and chronometer of my life. To Geddy. .-Vlex and .Veil (friends in my head): all the best! My heart, mind, and spirit is with you. even though 1 have never met you. nor am likely to. X ABSTRACT This dissertation contains multiwavelength observational projects on collisional ring galax ies. active galactic nuclei (AGNs). and neutral hydrogen observations of the collisional ring galaxy Arp 118 (an interacting pair NGC 11-14/1143). I have demonstrated in my dissertation research that: starburst knots of individual colli sional ring gala.xies cluster around the same, very young age suggesting that their formation was recently triggered by a coherent event all around the ring and strongly supports their pro duction as a result of the passage of a radial density wave: the atomic and molecular emission in the strongly interacting gala.xy .\rp 118 appears to be very asymmetrically distributed such that the .WV region of the gala.xy is mainly atomic hydrogen, whereas the SE region has an interstellar medium which is ~94% molecular hydrogen. One possiblity is that in the SE region there is a rapid conversion of HI to H). which may be a result of the highly off-center collision between .\GC 1144 and .\GC 1143: and the global far-infrared colors of Seyfert galaxies may be linked to their radio morphologies. I CHAPTER 1 INTRODUCTION 1.1 A Brief Description of Dissertation Projects and Organization of Introduction This dissertation is centered around detailed nuiltiwavelength projects on collisional ring galaxies and active galactic nuclei, and V'L.\ C-array HI observations made on a single colli sional ring galaxy. In particular. I present the results of a project designed to determined the abundances and star formation properties of a sample of S collisional ring galaxies. I report on the content and distribution of neutral hydrogen, and the relation between HI and CO emis sion. within the strongly interacting ring galaxy .\rp 118. and I describe the results of a pilot study which focuses on the possible relationships between active galactic nuclei or starburst activity and the morphology of the arc-second radio emission. The introduction is organized as follows. I briefly review (in Section 1.2) the morphological classification of galaxies. Since all three disaertatiuii projects contain e.xamples of interacting galaxies, and touch upon some aspect of star formation within them. I review (in section LI}) the star formation properties of interacting galaxies. .Additionally. I introduce ring galaxies, and de.scribc some of their basic properties. In section 1.4. I discuss active galactic nuclei (.\G.\s) and their basic properties. These reviews are not meant to be comprehensive. I am writing them in order to "set the table" by providing some basic background information. This dissertation contains many different forms of data: radio continuum and line, optical images and spectra, and infrared spectra and images. While the data reduction techniques are standard and adequately discussed within the chapters involving individual projects, there arc certain tools and techniques I have employed that deserve more complete treatment. Therefore, in .section 1.5. I give a detailed description of the interpretation of optical spectra. In particular. •1 the use of optical spectra to: i) determine the excitation mechanisms operating on emission- line gas (e.g.. if emission line gas is thermally, non-thermally. or shock heated), and ii) e.xtract metal abundances from the spectra of H II regions. Section 1.6 contains a summary of how the dissertation is organized. 1.2 Morphological Classification of Galaxies The most widely used classification scheme of gala.Kies was introduced by Edwin Hubble in 1936 (later updated and revised by Sandage in 1961). In its simplest form, three basic morphological types of galaxies are recognized: elliptical, spiral, and irregular. The Hubble sequence, or Hubble tuning-fork diagram, is shown in Figure l.I. .Along the handle of the fork are the elliptical gala.xies. becoming more highly flattened in appearance proceeding from left-to-right, .\long the top "tine" are the unbarred spiral galaxies, which progressively (again proceeding from left-to-right) have smaller bulge-to-disk ratios and larger, more prominent spiral arms which also become more open and less tightly wrapped. The bottom "tine" is completely analogous to the upper fork, e.xcept that the spiral arms end in a bar. which cross the nuclei of the gala.xies. .\t the branching of the fork are the SO galaxies, which have a combination of properties intermediate between the actively star forming spirals and the more quiescent ellipticals: SO galaxies are characterized by a smooth central spheroidal component surrounded by a generally featureless, flat, small disk. The convention is to call Es. SO and Sa "early-type" gala.xies. whereas Sc galaxies are generally called "late-type". Elliptical N«buki« Figure 1.1 Hubble Sequence, from Struve i: Zerbergs (1962). :} In the Hubble-Sandage system, there are galaxies that have an irregular or pecidiar mor phology. e.g.. the Large Magellanic Cloud (LMC) and .\I82. which do not readily fit into the morphological sequence. Magellanic-type irregulars are referred to as Type [ irregulars (Irr I), and very disturbed, highly peculiar gala.\ies (e.g. M82) are classified as Type 11 irregulars. Finally, this system is defined almost e.xclusively in terms of giant gala.xies. however, there e.xists numerous dwarf gala.xies (elliptical and irregular). .Another classification system was developed by G. de V'aucouleurs (19-59). In simple terms, the de V'aucouleurs system includes: e.vplicit reference to rings and s-shaped objects (by inclu sion of the symbols r and s. respectively): continued the spiral sequence beyond Sc to include gala.>:y types Sd. Sm. and 10 and Im (10 denote non-Magellanic irregular gala.xies. like .\IS2. and Im for Magellanic irregular gala.vies): and spiral gala.Kies. like those gala.xies on the top tine of the Hubble sequence, were designated S.\ (for e.xample. S.-Va. S.\b. and S.-Vc correspond to Hubble's gala.xy-types Sa. .Sb. and Sc). 1.3 The Star Formation Properties of Interacting Galaxies Gala.Kies, once thought to be "island universes", were believed to have their global proper ties (e.g. mass, morphology, star formation rates) imprinted upon them during the processes of their formation. However, modern understanding is that gala.xy-gala.xy interactions may play an important role in the development, and evolution, of the main structural properties of galaxies, [-'or example, the merger of spiral gala.xies to form ellipticals: tidal interactions giving rise to spiral (or radial) density waves: collisionally induced massive star formation, perhaps leading to rapid chemical enrichment (see Struck 199S for a comprehensive review of gala.xy collisions, and for further references). In a seminal theoretical paper on gala.xy-gala.xy interactions. Toomre 5c Toomre (1972: hereafter TT) used collisionless (star-like) test particle .simulations to reproduce the morpho logical features which are today generally accepted as being associated with interacting systems: tidal tails and bridges, .\bove and beyond what their simulations revealed about interactions amongst galaxies. TT further speculated that "...Would not the violent mechanical agitation 4 of a close tidal encounter - let alone an actual merger - already bring deep into a galaxy a fairly sudden supply of fresh fuel in the form of interstellar material, either from its own outlying disk or by accretion from its partner" (Toomre .'ic Toomre 1972). In other words, the "mechanical-agitation" of interactions may funnel gas into the liearts of gala.xies and provide the necessary fuel for a phase of enhanced star formation. Observational support for enhanced star formation rates in interacting gala.xies was pre sented by Larson &: Tinsley (1978: hereafter LT). where they compared the distributions of L - B and B - V colors of a sample of "normal" (e.g.. non-interacting gala.vies chosen from the Hubble .\tlas of Gala.Kies (Sandage 1961). hereafter called Hubble gala.\ies) and interacting gala.vies (e.g.. gala.\;ies in the .A.rp .-Vtlas) with a grid of gala.Ky models, each constructed to have different star formation rates (constant and bursts with various timescales). The burst models revealed that the colors of gala.>:ies evolve rapidly in the first Gyr. suggesting that there should be a large scatter in the colors of gala.xies with recent episodes of rapid massive star formation. Over-plotting the models with the observed colors of their sample of gala.xies. LT found that: i) the large .scatter in the colors of gala.xies from the .\rp .A.tlas were consistent with bursts of duration 2 x 10' to 2 x 10*^ years, and having involved up to about o% of the total mass of the gala.Ky and ii) Hubble gala.xies have colors con.sistent with star formation rates that are monotonically decreasing with time, which have not changed significantly over timescales > .5 X years. Bushouse (1987) found the global SF'Rs in a sample of strongly interacting disk-type gala.x ies were ~2.5 times higher than for isolated spirals, albeit with much overlap and scatter. Bushouse also concluded that when an interacting gala.xy experiences enhanced star forma tion. it preferentially occurs in (or near) the nucleus (Bushouse 1987). Kennicutt and Keel (198-1) investigated the effects of gala.xy interactions on the strength of nuclear activity in interacting spirals and found that, on average, these gala.xies had stronger nuclear hydrogen Balmer emi.ssion lines, and equivalent widths, than in a sample of non- interacting spirals. The level of ionization, and the existence of Seyfert nuclei, was al.so en hanced relative to the non-interacting sample (Kennicutt Keel 1984: hereafter KK84). Inter o estingly. Bushouse (1986) found fewer Seyfert galaxies in the nuclei of a sample of "violently" interacting galaxies. In a study of close pairs of galaxies and strongly interacting (Arp) galaxies. Keiiniciitt el al. (1987: hereafter K87) found systematically higher Ha and infrared emission, on average, than a control sample of non-interacting galaxies. The paired sample had a large fraction that exhibited relatively normal star formation properties, whereas the the .\rp sample was heavily biased towards high star formation activity. .Also. F\87 showed that strong disk line-emission is almost always accompanied by strong nuclear emission. These examples from the literature demonstrate that one must be cautious: the fact that some interacting galaxies show enhanced star formation does not imply that all do. Sta.tistica.lly speaking, tidal encounters often do lead to rapid massive star formation but. this isn't always the case. Bushouse (1986). in an optical spectrophotometric survey of the nuclei of violently interacting pairs of galaxies, found a wide range of optical emission-line strengths. Some nuclei contained vigorous massive star formation, while about a third of his sample contained weak or no detectable optical emission and were characterized by having absorption lines indicative of either an A star population or an old. elliptical gala.xy-like .stellar population (Bushouse 1986). Therefore, a third of his sample displayed no evidence for significant star formation on timescales > 1 Gyr! special cla.ss of interacting galaxies near and dear to me are the collisional ring galaxies. Ring galaxies are believed to be formed when an intruder strikes, in a nearly bulls-eye collision, the center of a larger disk system driving an outwardly expanding radial density wave (Lynds ic Toomre 1976 (see Figure 1.2); Theys &: Spiegel 1976. 1977: Toomre 1978: see .\ppleton 1*0 Struck-Marcell 1996 for a comprehensive review of ring galaxies). One consequence of the passage of the density wave is triggering of massive star formation, leaving aging star clusters in its wake. Ring galaxies are beautiful and spectacular examples of interaction induced massive star formation. These systems are useful for studying massive star formation for a number of reasons (I will name just a few). In the nuclear regions of galaxies one is often confronted with Figure 1.2 Formation of a ring galaxy, from Lynds and Toomre (1976). the problem of disentangling multiple sources of ionizing radiation, for e.xample distinguishing between thermal emission from hot stars and a power-law continuum generated by a central engine. In an e.xtended ring wave, massive star formation can be studied without the possible complicating effects an AGX. Stellar clusters, left in the wake of the ring, can be used to follow, in detail, the evolution of their luminosity (e.g. Appleton 1998: .-Vppleton .lacobs 1998). Rings waves can be used to study the hydrodynamic response of a disk to a perturbative, non-disruptive, encounter (e.g. Gerber. Lamb, Balsara 1996: Struck-Marcell Higdon 1993: .Vppleton Struck-Marcell 1987a.b). and to study the relationship between gas density and stellar density waves to test models of wavc-induced massive star formation (.Marstoii i: .-Vppleton 199-5; Charmandaris. .Appleton. &: Marston 1993: Appleton. Schombert. cV: Robson 1992). 1.4 Active Galaxies The first hint that the nuclei of some galaxies were different from others was first recognized by the discovery of emission lines, first by Edward Fath (1909) and then by V. Slipher (1917). in .\GC 1068. Edwin Hubble (1926) noted "planetary-nebula-type" emission lines in the nucleus of .\GC 1068. 1051. and 4151. Carl Seyfert (1943) presented evidence that a small fraction (1 gala.KV in 10"') of galaxies contained, in their very bright "stellar-like" nuclei, and exhibited high I ionization emission lines. The strengths of the lines, in wliat are now referred to as "Seyfert" galaxies, suggested the e.xistence of energetic phenomena that cannot be readily explainable in terms of a typical stellar population. For example, in a typical Seyfert galaxy, the total energy output at visible \va%-elengths is comparable to the total energy output of all the stars in the gala.xy (~10" L,; ). Other galaxies were identified as having unusually strong radio sources in their nuclei, possessing large radio jets and lobes that extend far beyond the optical image of the galaxy (e.g.. Cygniis and Centaurus .-V: these are examples of a class often referred to as radio-loud .-VG.\). class of "radio-quiet" .-VG.\ was also discovered (not as radio-luminous as the sources possessing the kpc sized radio-lobes, but certainly still "radio-loud" with respect to normal field galaxies). Also, a fraction of these radio galaxies had a rather stellar-like appearance, with no sign of a galaxy, nor any nebulosity. These "quasi-stellar objects" had continuous spectra, with no absorption features, but contained broad emission lines (these objects are now referred to as quasars, or QSOs). Today we recognize that the galaxies described above (e.g. Seyferts and radio-loud and radio-quiet AGXs. and QSOs) are a "unified" class of gala.Kv collectively called active galactic nuclei, or .-\G.\s. The.se gala.xies are unified, in the sense that all .-VG.Vs are believed to be powered by the accretion of matter onto a super massive (10'' to lO'"' .\[ ..) black hole. Accretion onto a black hole is capable of producing copious amounts of energ\' due to the gravitational collapse of matter, and the subsequent release of energy. Simply put. this is a process of conversion of mass to energy with some efficiency q. such that there is a relea.sc of energy E from the accreting mass \[. given by E = q\lc-. The luminosity is L = dE/dt = f/.Mc^. A modest efficiency of 10% and small accretion rate (.\l) of ~l M. is enough to account for the large observed luminosities of .-VG.N's. ~10^' to lO'^ L... .\ctive galaxies can. however, span a wide range of luminosities and phenomena (i.e.. not all .\G.\'s have kiloparsec sized radio lobes) from the quasars, which are the most distant and most luminous type of .\G.\. down to the ubiquitous, more-quiescent LINERs (a class of low luminosity active gala.vy characterized by low-ionization specics in its spectra; this type of activity is discussed later). 8 What are the (optical) spectral characteristics of these galaxies? Seyfert galaxies have two main types of emission-lines: i) very broad (P'WHMs of order a few x 10^ km s~') permitted lines of H and He and ii) narrower (FWHMs of order 10^ km s~') forbidden lines, like [OIII] AA4959.5007. [MI] AA65 48,6584. and [Sit] AA6717.67:U. Some Seyfert galaxies have only narrow lines (both permitted and forbidden), and others have a mixture of broad permitted and narrow forbidden lines. The spectra of Seyfert galaxies can sometimes show broad permitted hydrogen lines superimposed on a narrower component, in addition to the narrow forbidden lines. If a Seyfert galaxy spectrum contains only narrow-lines it is referred to as a Seyfert 2 gala.xy. if it contains broad and narrow lines it is referred to as a Seyfert I gala.vy. and if it is a mix of broad and narrow permitted lines it is a Seyfert l.o galaxy. Radio galaxies come in two spectral types: those with broad-lines and those with narrow lines (broad-line radio galaxies. BLRGs: narrow-line radio gala.xies. .VLRGs). BLRGs have spectra that resemble closely the Seyfert 1.5 spectra, whereas the NLRGs have spectra that are nearly identical to Seyfert 2 spectra. .-V word of caution about Seyfert types is warranted. The broad-line components in the spectra of some Seyfert galaxies are highly variable, such that the broad-lines can become very weak and at other times become very prominent (e.g. XGC 5548: Peterson k. Wanders 199S). Hence, at times a Seyfert 1.5 gala.xy may look more like a Seyfert 1. and other times look more like a Seyfert 2. Penston and Perez (1984) reported the broad-lines in the Seyfert 1 gala.xies .\GC 4151 and ;}C390.."} nearly completely disappeared, such that these galaxies appeared to become Seyfert 2 galaxies (the broad lines eventually returned). The mechanisms that create variability in these lines remains controversial and is poorly understood. In the "standard model" of an .\GN'. an accreting super massive black hole is a non thermal. power-law source of photons which encounter two distinctive regions: the small and dense "broad-line region" (BLR: R = ~10~^ pc. Ue = ~10^ to lO'" cm~'^. and T^ = ~10"* K) surrounded by a larger, and less dense, "narrow-line region" (.\LR: R = ~10^ pc. Ue = 10^ to lO' cm" ', and T^ = ~10"' K). Immediately surrounding the BLR is an opac|ue. dusty molecular torus such that the differences between the Seyfert classes are simply a function of orientation 9 (see Figure The dust torus may collimate the escaping high-energy photons resulting ia the observed "ionization cones" (e.g. Pogge 1988). Ionization cones are X shaped regions of high ionization and excitation, seen mainly in Seyfert 2 galaxies, where X. in essence, "marks the spot" of the location of the .-VG.\. The dust torus is also believed to be the location of significant FIR emission, which would arise from the re-radiation of dust heated by the AGX. lonizadon zone boundary Gas & dust 'torus' NLR clouds (50-100 pc) ENLHO clouds n (kpc)/ 'Seyfert V Figure l.;j Conceptual scheme of the unified model of an AG.V. The black dot at the center is the BLR. The arrows represent an observers viewing angle of the .-VGX. From Peterson (1997). How ubiquitous is star formation in the circumnuclear (i.e. < 1-2 kpc from the nucleus) regions of galaxies that host an .-VG.V? Recall KKS4 presented evidence their sample of in teracting galaxies, in addition to having enhanced nuclear star formation, contained a large fraction of Seyfert gala.xies. Kennicutt. Keel. Blaha (1989) compared the physical conditions of nuclear, hot spot ("hot spots" are rings of star formation which coincide with inner Lindblad resonances), and disk H II regions and found that nuclear H II regions have anomalously strong 10 forbiclcleri-line emission, consistent with a second source of ionization; either from strong shocks or an active nucleus. There is evidence that the most luminous far-infrared (IRAS) galaxies may be either be powered by star formation (Soifer et al. 1987). luminous buried AGN's (Boyce et al. 1996). or both (e.g.. .Miles et al. 1996). Further observational evidence that .-VG-Vs are surrounded by m^issive star formation can be found in Heckman et al. 1997: Genzel et al. 1995: Baum et al. 199.}: Condon et al. 1991: Shields Filippenko 1990: Wilson 1988. 1.5 Interpretation of Optical Spectra 1.5.1 Excitation Mechanisms Operating on Narrow-Emission-Line Gas Building upon known classification schemes of H II regions which were based on single intensity ratios of emission lines, like [OIII] A5007/Hii. Baldwin. Phillips and Terlevich (1981. and references within: hereafter BPT) developed a family of two-parameter line ratio diagnos tic diagrams, in order to determine which set could best distinguish between the three key mechanisms at work in narrow-emission-line gas: i) photoionization by hot (OB) stars, ii) pho- toionization by a non-thermal, or power-law. continuum, and iii) shock heating. BPT compiled a large list of optical spectra for different classes of narrow-emission-line objects; planetary nebula, H II regions (both galactic and extragalactic). .\LRGs and Seyfert 2 galaxies, and the LI.N'ERs (LL\ERS are discussed more fully below). They noted that the following combina tions of emission lines, plotted against each other, were successful in separating these objects from each other: a) [01] A6.'{00/Ha versus [Oil] A;5727/[Oni] A.5007. b) [OIII] Ao700/HJ versus [-\TI] A6.584/Ha. and c) [01] A6300/[0111] A.5007 versus [Oil] A;3727/[Oni] A5007. Proper corrections for extinction effects are clearly a concern, as are abundance variations among the various objects studied. For example, the use of the [Oil] A.'}727/[0I1I] Ao007 lint^ ratio is very sensitive to uncertainties in reddening corrections, due to the largo separation in wavelength between the lines. Furthermore, for any observational line-ratio diagnostic to have physical meaning, it should agree with theoretical models which predict the line-ratios of gas excited by different mechanisms, \eilleux Osterbrock (1987: hereafter V'O) undertook a similar project to BPT. choosing line ratios that were closely spaced in wavelength to reduce 11 errors in reddening corrections. Predicted line-ratios, based on models of the three excitation mechanisms list above, were compared with observed line-ratios from an expanded database of optical spectra of extragalactic H II regions, starburst galaxies (or H II region-like gala.xies). Seyfert 2 and .VLRGs. and LINERS. Various abundances, electron densities, ionization param eters. and shock velocities were explored in these models. The line ratio diagrams used by V"0 were: [Nil] A6o83/Ha versus [OIII] A5007/HJ. [SII] AA67I6.6731/Ha versus [OIII] A.5007/HJ. and [01] A6300/Ha versus [OIII] A5007/H.J. In Figure 1.4 I reproduce the V'O diagrams (taken from Osterbrock (1989). Chapter 12) which revealed, as did the BPT diagrams, these line-ratio diagnostics result in good separation of H II region-like gala.xies from .\GNs. .Although not shown in Figure 1.4. \'0 attempted to fit shock models to their data, and concluded, from the generally poor agreement with their data, that shock heating is not a dominate excitation mechanism in .-\.G.Ns(V'0). This conclusion has been challenged (cf. Dopita Sutherland 1995. see below) by improved shock models that can. in fact, reproduce the line-ratios observed in some .-\GNs. BPT and V'O demonstrated that it is possible to classify narrow-emission-line galaxies in a quantitative, physically meaningful way. However, even with these powerful tools, one must be cautious. There are examples of galaxies that can change their classification, depending on the diagnostic diagram used. These galaxies generally have spectra that place them on the boarder between AGN/I.INER, AGN/HII region-like, or L1.\ER/H II region-like emission. Such gala.xies may be "transition" or "composite" classes, with multiple e.Kcitation mechanisms at work on the gas. There are AG.Ns that have been shown to have vigorous massive star formation on scales of < 1 kpc. like .\GC 7469 (Genzel et al. 1995) and Mrk 477 (Heckman et al. 1997). .\lso. controversy still surrounds the origin of LI.XERS. e.g. shocks versus photoionization. and in the next paragraph 1 discuss these object more fully. Ll.VERS (cf. Heckman 1980) are a class of active gala.xy characterized by strong low- ionization species, such asO". S"*", and .N"*". and weaker high-ionization lines. The low ionization lines observed in Ll.VERS are still relatively strong compared to permitted recombination lines and have widths comparable to Seyfert 2 and NLRGs. Further. LINERS occur frequently 12 i «• O NO 0«J u o >•< -u - -u -u U|0>t Q AM) { I I I 1 I I : I I I I i I I I I I s u - ) Oo. ' ?0 -05^ 'IJ -1.0 «4S U|(B B] OTU 4-uniVOb USOft TP -«8^ o O \ -IJ -10 -U -20 -u -U -*3 «• i«(foaajeovat»i«Mi) Figure 1.4 Narrow-emission-line-ratio diagnostics. The solid line is the nominal dividing line between H 11 region-like galaxies and AGXs. The dashed lines are pho- toionization models. Photoionization models from hot OB stars, not shown in these figures, provided excellent fits to the H ll-region-like galaxies. From Osterbrock (1989). i:} in normal galaxies, particularly late-type galaxies (Heckman 1980). Heckman (191^0) showed that, in a plot of [01] A6300/Ha versus [Oil] A;i727/[OIIl] AoOOT. Seyfert galaxies and LIXERS are (smoothly) separated, but with some overlap. Heckman (1980) claimed that the line ratios of LI.\ERS were best fit by shocks, with velocities of the order ~10- km s~'. and that the differences between Seyfert galaxies and LI.NERS was one of the relative importance of photoionization vs. shock heating. The origin of the [01] AG300 line in .\G.\s was argued, in \'0. to be predominantly produced by photoionization and not enhanced in .-VGNs as a result of shock heating. This conclusion was based on poor shock model fits in the [Oil] Ay727/[OIII] Ao007 versus [01] A6300/Ha line ratio diagram. However recent shock models (e.g. Dopita Sutherland 1995) better match the observed line ratios of Seyfert gala.xies and cast doubt on this conclusion. Another interesting possibility for the origin of LIXERS are shocks produced by nuclear starburst super winds (Heckman. .\rmus. Miley 1990). Realistically LIXERS are probably powered by a number of different mechanisms, and the relative importance of shocks and photoionization must be con.sidered on a gala.Ky-by-gala.xy manner. 1.5.2 Metal Abundances in H II Regions The purpose of this section is to describe how to derive o.xygen and nitrogen abundances from the optical spectra of H II regions. I am going to concentrate on the "nuts-and-bolts" of the procedures involved instead of presentifig a comprehensive review the astrophysics of gaseous nebulae, or of the importance of abundance determinations. Having said that. I would like to stress that determining reliable abundances (either nebular or stellar) is vital to our understanding of chemical evolution in gala.xies. deeper understanding of chemical evolution, in turn, provides essential clues which may shed light on gala.xy evolution. I refer the interested reader's attention to the comprehensive review of chemical evolution in nearby galaxies by Evan Skillman (1998). In the following two sections I outline the procedures used to calculate ionic abundances (the "direct" method) and empirical abundances (the "bright-line" method). Naturally, the first step in the determination of abundances, whether it be via the direct or bright-line method. 14 requires the extraction of line-fluxes from spectra, their subsequent corrections for reddening, and a determination of errors in the line fluxes. Section 2.4.1 in Chapter 2 contains a thorough review of these procedures, which need not be repeated here. 1.5.2.1 Derivation of Ionic Abundances or the "Direct" Method In order to derive ionic abundances, high quality spectra must be obtained. These spectra must have a sufficient number of key emission lines in order to derive key physical parameters such a.s electron temperature and density. The high e.xcitation [0111] emission-lines (A4363 and AA49o9.5007) are a good probe of the electron temperature of a nebula, and the [SU] doublet (AA6716.6731) is a good indicator of electron density. I refer the interested reader to Osterbrock (1989. in particular see Chapter 5) for a complete treatment of the physics, and procedures, behind the temperature and density determinations in gaseous nebulae. Once the electron temperature has been determined the abundance of an ionic species, relative to singly ionized hydrogen, may be computed using: -V(.V+') ^ /(A) .V(W+) l[H3) j[\) • ^ ' where .\ is the abundance of an ion. .X"*"': I(A)/I(H.i) is the observed, reddening corrected line ratio; and j is the emissivily of the line. The emissivity of a collisionally excited line, when collisional dG-e.xcitation is negligible, is given by: m = ,v, ,v, ;;,,.2, T 7 2 y where .\', is the number density of the ion of interest. .V is the electron number density, h is Planck's constant, c is the speed of light. A is the wavelength of the transition, fi is the average electron collision strength from the lower level, g is the statistical weight of the lower level, k is Boltzman's constant. T the electron temperature, and AE is the energy difference between the two levels. 15 Assuming that the hydrogen is completely ionized such that H""" = H. the ratio of an ionized species relative to ionized hydrogen. .\'(X'^')/N'(H'^). becomes N'(.X.'*"')/N'(H). The total abundances a species is. then, simply a sum of the individual abundances of an ionized species: (1-3) However, some (usually higher) ionization species aren't always available in optical spectra, therefore relying on optical lines alone may underestimate the true abundances. For e.xample. the presence of O"'"''"'' and N'*'"'' (infrared emission-lines) may contribute significantly to this sum. Often the presence of higher ionization species is corrected for by boot-strapping fluxes of lower ionization species to obtain the fluxes from higher ionization species. For e.xample. in order to compute the total nitrogen abundances, one must correct for the presence of XllI by scaling up the computed XH flux using .XHI = (OHI/OII).Xn. Therefore, we adopted: •V(-V) .\{oif -f- oiif) .\'(.\'fn (l-l) \iOf[) .\{H) • This is generally a reasonable approach because .XII and Oil have similar ionization poten tials and XT and 01 have ionization potentials that are almost identical with HI. This is not as accurate a procedure, say. as model fitting spectra or actually ublaiiiiiig spectra in oilier wavelength bands to directly compute the missing line-flu.xes. Further, relying on similarities between ionization potentials, while approximately a good guide, isn't always so: see the cau tionary comments in Garnett 1990. Despite these uncertainties, ionic abundances are generally accurate to 10-20% (Skillman. 199S). 1.5.2.2 Empirical Abundance Determinations or the "Bright-Line" Method The temperature sensitive line [OIII] A4363 is very weak and is not always detected. How can abundances be calculated if the physical conditions (especially the temperature!) within a nebula are unknown? .\s Pagel et al. (1979) showed, it is possible to determine oxygen abun dances when a few key ""bright-lines" exist in a spectrum. The basic assumptions that Pagel 16 et al. made, in order to relate the oxygen abundances to these key spectral lines, were: i) the temperature of an H H region is governed by the balance between the heating, due to a central exciting star (or stars), and cooling, which is dominated by coHisional excitation of singly and doubly ionized o.xygen. and ii) a "universal" relationship between (oxygen) abundance and the effective temperature of the exciting stars (Pagel et al. 1979). The upshot of this is that: i) higher o.xygen abundances result in lower the nebular temperatures, and vice-versa, and ii) the higher the metallicity (i.e. o.xygen content) of an H II region, the lower the effective temperatures of the e.xciting stars. .-Vt lower metallicities the majority of the cooling in H II regions is from the high excitation optical [OIII] lines, while at higher metallicities the cooling is dominated by the infrared fine- structure [OIII] lines. L'sing the above assumptions, and the realization that the strengths of the [OIII] lines (A;i727-|-49o0-i-o007) relative to HJ (the .so-called R23 index) increa^ie with decreasing o.xygen abundance. Pagel et al. developed a calibration of the R23 index ratio to the total o.xygen abundance for H II regions. Edmunds k. Pagel (19S4) latter re-calibrated this relationship, and other investigators have developed their own calibration of the R j.-} index with o.xygen abundance (e.g. Dopita Evans 1986; .\lcCall. Rybski. Shields l98o). Figure 1.5 contains a plot of the these three calibrations, along with a fourth by \'ilchez (19S-S). .Notice that at low metallicities the empirical relationship is double valued. This stems from the fact that the R>3 index cannot continue to increase indefinitely as the oxygen abundance decreases. One must keep in mind that when this relationship is double-valued, there is an ambiguity between a hot. o.xygen poor H II region and a cooler, more metal rich H II region. One way to determine which branch of the curve is applicable, determine the strength of the optical [.MI] or [SII] emission lines, which are much stronger at high o.xygen abundances and much weaker at lower abundances. The R23 index can be used, in conjunction with the various calibrations, to compute an o.xygen abundance. In the mean, the three (Edmunds Pagel 1984: Dopita L Evans 19S(j: .\IcCall. Rybski. Sc Shields 1985) methods may be approximated by 12 -|- log(0/H) = 9.265 - 0.33x - 0.202x" - 0.207.x"^ - 0..'}.J.i.x' where x = log(Ro3) (Zaritsky. Kennicutt. Huchra 199-1. 17 LOG 0/H +12 •— (SUN) M83 NUC ORION NGC 804 IZW1B —2 0 •2 •4 •6 •8 10 LOG[([Oll]+jDlllj)/Hp] Figure l.o Relation between R>3 index and oxygen abvindance for H II regions. Tiie dashed line is Edmunds lI: Pagel (19S-1): the dotted line McCall. Rybski. Shields (lOcSo): the dash-dot line Dopita 1.6 Dissertation Organization This dissertation consists of five chapters. The first chapter is the introduction. Chapter 2 is a paper on the metallicity and star formation properties of starburst knots in collisional ring galaxies, [n the third chapter I present the results of neutral hydrogen observations of the strongly interacting, ring galaxy .\rp 118. Chapter -1 is a paper consisting of radio images and optical/near-iafrared spectroscopy of 12 southern Seyfert galaxies. The fifth chapter contains a general summary of the major conclusions, and a brief description of future/ongoing research plans. A list of references for Chapters 1 and .j is included at the end of the dissertation. 19 CHAPTER 2 MULTIWAVELENGTH OBSERVATIONS OF COLLISIONAL RING GALAXIES. III. OXYGEN/NITROGEN ABUNDANCES AND STAR-FORMATION PROPERTIES OF RING KNOTS A paper submitted to The Astronomical Journal M. A. Bransford. P. N. Appleton. A.P. Marston 2.1 Abstract This is the third paper in a series examining the multiwavelength properties of a sample of northern ring galaxies, in which we present optical long-slit spectra and broad-band B\"R and .]HK colors of individual star forming knots embedded in the ring. We present tiie oxygen and nitrogen abundances of the starburst knots, compare the reddening corrected colors of individual star clusters with recent models of stellar evolution, and present kinematic evidence that one system. LT -U. has a simlutaneously expanding and rotating ring, consistent with ring galaxy models that predict an outwardly expanding density wave. Ring galaxies are found to have sub-solar nietallicities in the range one-half to one-fifth solar in [0/H] and ratios. There is a suggestion of an increase in the mean nitrogen abundance for the rings of larger linear size, but that the o.xygen shows no trend with ring diameter. Ignoring the pos.sible effects of chemical enrichment in the ring waves, the oxygen and nitrogen abundances are consistent with what one would expect for a range of Hubble-types. Interestingly, the nietallicities are consistent with a morphological segregation of Hubble-type with ring size: the smaller rings inhabit smaller, late-type galaxies whereas the larger rings inhabit larger, mid- to early-type galaxies. 20 Comparison of the colors of individual starburst knots with recent numerical models of stellar evolution revealed that the knots have ages that lie in the range from 4 to 80 Myr. These ages are less than the dynamical ages of the rings, which are typically 100 to ;}00 NLyr. The fact that the knots from individual ring galaxies cluster around the same, very young age argues that their formation is recently triggered by a coherent event all around the ring. This strongly supports the triggering as being due to the propagation of a radial density wave. 2.2 Introduction Ring gala.Kies are a class of interacting, extended starburst galaxy believed to be formed when an intruder strikes, in a nearly bulls-eye collision, the center of a larger disk system driving an outwardly expanding radial density wave (Lynds Toomre 1976: Theys Toomre 1979: see .-Vppleton (S: Struck-Marcell 1996 for a comprehensi\e review of ring gala.xies). One consequence of the passage of the density wave is it triggers massive star formation, leaving in its wake aging star clusters. In previous investigations of ring galaxies. Appleton .Marston (1997. Paper 1: hereafter .\.\[) presented the details of a sample of northern ring galaxies, their global BV'R and JHK colors, and reported the existence of radial color gradients in some of the larger rings (see also .\Iarcum. .-Vppleton. L Higdon 1992). .Marston .\ppleton (1995. Paper II; hereafter .\I.\) probed the global massive star formation of the sample via broad-band Ha images. These images were used to determine star formation rates, the distribution of new- stars. and the relation between the gas density and stellar density waves to test models of wave-induced massive star formation. Other, more detailed studies have been published for .\rp 10 (Charmandaris, .\ppIeton. Marston 1993: Charmandaris .-Vppleton 1996) and \'I1 Zw 466 (.-Vppleton. Charmandaris. (S: Struck 1996). We continue our examination of the multiwavelength properties of this sample. We present optical long-slit spectra and broad-band colors of individual knots embedded iu the rings. We use this data to answer the following ciuestions: (1) What are the elemental abundances? (2) Do the abundances correlate with ring size? (3) .Are the star clusters younger than the dynamical age of the ring. i.e.. the time taken for the ring to propagate to its current size? 21 (-1) Is there any evidence that the knots were created by the passage of a radially expanding density wave? In §2.3 we discuss the details of the observations and data reduction. In §2.4 we describe the methods used to extract oxygen and nitrogen abundances and the BV'R and .IHK colors of the knots. In §2.5 we discuss the results, and implications, of the o.xygen and nitrogen abundances derived for the knots, discuss the Balmer absorption lines observed in the spectra of II Zw 28 and .A.bell 76. In §2.6 we discuss the kinematics of the ring galaxies. In §2.7 we compare the colors of the knots with the population synthesis models of Leitherer ic Heckman (1995; hereafter LH95) and Bruzual Chariot (1995: hereafter BC95). In §2.7 we present our summary and conclusions. We assume distances based on Hq = 75 km s~' Mpc~'. 2.3 The Observations The optical spectroscopic observations are described in this section. Details concerning the optical and near-infrared CCD imaging can be found in .-VM. Optical long-slit spectra were obtained for the sample at Kitt Peak .National Observatory (KP.VO) on two different observing runs: 29-;}0 .lune/l .July. 1992. and 2>!-;{0 .November, 1992. The observations were made with the 2.1-meter telescope, in conjunction with the Gold Spectrograph CCD camera (GOLDC.VM), using a FORD 3K x IK CCD detector. Details on GOLDC.VM may be found in the Gold Camera L'ser's .Manual (De \euy. Carder k Harmer 1995). With the Gold Spectrograph we used a grating with 500 lines mm"' which was blazed at 5500 This provided a 1.52 .\ pi.xel"' spectral and a 1.54" pi.xel"' plate scale. Typical seeing conditions varied from 1-2". Slit widths were set to 2" during observations of the ring galaxies. Observations of the ring gala.xies, across several slit-positions, were accompanied by spectrophotometric observations of standard stars to allow for flux calibration. Wide slit widths were used (6") for the observations of standard stars. Integration times for the ring galaxy spectra ranged from 2300-3000 s, with 3000 s being typical. The integration time for the standard stars was nominally ~100 s. The IRS standards observed were: HD 2S57, HD 74721, Feige 25. and Feige 34 (Barnes Hayes 1984). Pairs of Helium-Neon-.\rgon arc lamps 22 were observed, bracketed with the galaxy observations, to allow for wavelength calibration. At the beginning of each night, bias and dark frames were taken in conjunction with observations of quartz lamps to allow for flat-fielding of the images. The data reduction of the spectra was completed in the standard way. using the long-slit packages in IR.\F. Table 2.1 contains the details of the optical spectroscopic observations. 2.4 Determination of the Oxygen and Nitrogen Abundances and Colors of the Knots 2.4.1 Extraction of Spectra and Determination of Line Fluxes We present in Figure 2.1 grey-scale B-Band images of the rings overlayed with the positions of GOLDC.-VM slit positions, along with identification of the knots for which wo have performed broad-band photometry. The photometric software apertures used in the extraction of the photometry are also shown as circles on the figures. We describe the details of the photometry in §2.3. In Figure 2.2 we show representative spectra of some of the knots, which are identified within the Figure. Knots where extracted from the spectra by setting apertures within the 2-d spectral images using the IR.VF routine APSLM. We emphasize that these are spectra of individual knots and not. for example, global spectra of the entire gala.KV. We used the SPLOT routine in IR.-VF to measure, and deblend. the emission line flu.xes. Flu.xes were corrected for internal reddening relative to using: ,oC(WJ)/(.\) 10^(2.1) I(HJ) F{FI3) where 1 is the true dereddened flux. F is the observed flux. f(A) is derived from the standard interstellar reddening curve for the Gala.xy from Savage Mathis (1979). and C(H.:^) is the logarithmic extinction constant which is calculated from the observed Hq/H J ratio. Here we have a.ssumed the relative line emissivities for case B recombination from Osterbrock (19^9). Uncertainties in the line flux ratios, relative to H.i. were calculated using the formula given by Skillman Kennicutt (1993). That is. the fractional error in the line flux ratio can be calculated using (where numbers are in accumulated electrons); 23 — — {C'l + C'> -f [rio/\/n^ S + n.4.\" + {•2.:if{X)6C{HJ) Lf + (O.Ol/:)- + (2.2) where SL/L is the fractional uncertainty in the line strength (L). C'l is the total number of counts (line plus continuum and sky). C ) is the number of counts due to the continuum. S is the number of sky counts in a single row. no is the number of rows averaged for an object, n^ is the number of rows averaged for the sky. and n.AN" represents the readout noise (n is the number of integrations. A the area summed over in pixels, and .\ the RMS readout noise). The fifth term represents the uncertainties involved in dereddening the line fluxes, where is the uncertainty in the logarithmic extinction constant. The si.xth and seventh terms represent, respectively, errors in flat fielding (nominally about 1%) and errors in the fit to the sensitivity function used to flux calibrate the spectra (where d'F is the RMS to the fit which was nominally Fractional uncertainties varied from spectrum to spectrum and line to line, but is generally in the range 5-20%. with 8-10% being typical. Table 2.2 contains the strengths and uncertainties of the undereddened lines. Table 2.3 contains the strengths and uncertainties of the reddening corrected lines used to calculate the nitrogen and o.vygen abundances. We note the presence of very strong Balmer absorption features in the 11 Zw 28 spectrum (first noted by Sargent (1970)) across the northern half of the ring (which is collectively knot 1). and in the spectrum of Abell 76 for knot 2. These spectra arc shown in Figure 2.2. In order to deblend the Balmer emission lines from the absorption features (especially for H.i and H-, ). we have used a multiple Gaussian fit to the spectra. .\n example of this is also shown in Figure 2.2. The emission line flu.xes (particularly H.i) obtained from this procedure are less reliable than if there was no underlying absorption, with typical uncertainties of the order of 1-5-30%. However, as discussed later, we find consistency between the calculated values of .-Vv found using the ratios of Ha to HJ and H-/ to HJ. This gives us confidence that the fitting procedure is relatively reliable. 24 Figure 2.1 B-Banci grey scale images of the sample of ring galaxies overlayed with the positions of the slits used to obtain the optical spectra. In addition, the star forming clusters used to obtain color information have been circled and labeled. In these images north is up and east is to the left, e.xcept for .A.rp 10 which has a compass to indicate north and east. 25 Figure 2.1 (Continued) T 1 1 1 1 1 1 1 T "T I I I I I I I I I I I I I I I I R "1—I—I—I—I—r 20 1- sID N CO [OU] X3727 O o [Nelll] \386g Ca II H + He II Zw 28 [on] X3727 I Ca IIK to H12 through Hy 1 L-l I I I I I 1 I 1 I I I I I I L J I I I I I I I L J I I I L 3800 3900 4000 4100 4200 4300 4400 4500 \(A) I'igiirc 2.2 l{o|)rc.s('iilaliv(' oik- diiiuMisioiial .spectra of iiulividnal li II ri'^^ioiis williiii lli(> riii^.s. riic.sc arc spectra of individual knots and arc not, I'or cxaiii|>lc, global sju'ctra of the entire galaxy. Also shown arc the multiple (lanssian (its in the 11 Zw 2iS and .Mk'II 70 spectra. [Oil] A3727 Abell 76 knot 2 I in 4000 4500 5000 5500 6000 6500 7000 n u00 v n I I I I I I 1 I I I I I I I I I I I I I o Abell 76 -I tH Ca UK Ca 11 H + He [Oil] X3727 H12 through Hy J I L J I I I i I I I I I I I I I I I I I I I I I I I I I I I I U 3800 3900 4000 4100 4200 4300 4400 X(A) 2.2 ((Vmliiuu'd) Fx (x 10 ergs cm"^ s ^ A 0 ro 00 o o ro CO cn 1 t M 1 t i 1 i i I M 1 i 1 I i I t 1 1 1 4^ oO [on] X3727 o [on] X3727 o o o cn o cn4^ o 3= cn o o o Hp ts: [om] XA5007.4959 [om] XX5007,4959 'iinJ>® cn N—^ o o Oio o o oCJ> o PT CO o 3 Q O O CO GO a> cn cn o o oicn o •Nl o o o-Nl o o [sn] XX6716,6731 [sn] XA6716.6731 ( 1 I i I I M I 1 t I I I 1 1 I 1 1 1 t t 8(. Fx (x 10 ergs cm ^ s ^ A O W 03 I I i 1 I I i I I I I I I I I I 1 I [on] X3727 [on] X3727 o o o oU1 o ocn o H/3 o [on] X5007 [om] X5007 cnOl o o s O) a o o + > — o 33r o •0 z O) o _ > - >> a —• cnOi cn . o CO o o _ Ol 09 •1^ [sn] XX6716,6731 o-Nl [sn] XX6716.6731 o o I I I 61 Fx (x 10 " ergs cm ^ s ^ A [on] X3727 o o [on] X3727 — o ai o o cn 3=. o o o rc n:. [om] XX4959,5007 XX4959,5007 [on] cntn o a> o o s 2 x Q < to 3 + O N w S 'z ro o> o> cn 03 - o >- o > 09 GD — 09 cn — -Nj CB o - o o [Sn] XX6716.6731 J I > 1 '111 I I I I J I L OJ: 31 9.00E-16 II Zw 28 8.00E-16 Hp + abs 7.00E-1S 6.00E-16 [OIII] A.4959 5.00E-16 4.00E-16 3.00E-ie 4900 5000 5100 5200 Uaveiength AbeU 76 Hp +abs 6.00E-16 5.00E-16 [OIII] A.4959 4.00E-ie 4850 4900 4950 5000 5050 5100 Wavelength Figure 2.2 (Continued) TIIMI' 2.1 'L lic Saiii|)k' and Optical .S|H'I Iro.scopic Ob.sorvatioii.s Name Dale Slit Po.sitioii" P.A ( i;r 11 •2K Nov 92 IM 90 :i:i90-7()U) :i000 2K Nov 92 P2 90 ;i;j90-7()i0 ;jooo 29 Nov 92 IM 0 ;j;{90-7()i0 :{0()0 29 Nov 92 PI 0 ;{;{90-7(ji0 ;{0()0 29 Nov 92 P5 0 a;}90-7()i0 ;}000 NCiC ".Si'c I'imiri' I for position of slilh iiroiiiicl ilu' riiin- wv lisi only tlioM' slii positioiih tlwvt svi'iv visnl lo i-xlrm t al)iiii(laii('(' iiiforinat ion. 'Tiibli'2.2 Uiicoriwlcd Lino St ri'iigt lis /(|O///].\AI!J,5!J,,'i007) Naiiio Slit Position" 1(11/:')'' /(///') LT -11 PI l.iSiO.Of) 3.5;j±0.25 2.G()±().2;J l.()9±0.il IM ().(j7±0.()7 r).;}7±().i2 2.:Kj±0.2(j 2.03±0.19 P2 1.2N±().0!J ;j.S7±().29 l.()9±0.1() 1.3r)±0.11 P2 l.lliO.OS i.i(}±().;n 2.:{2±().12 1.8;}±().11 0.1()±0.07 2.70±0.2l i.9;}±().2i 2.oi±i).:») PI ().!)!)±().l() 2.12±().28 2.22±().2r) i.()i±{).i:i P5 ().r)i±().()r) ;{.81±0.2H 2.i2±().i;{ i.;{(j±o.i;{ P5 o.fjoio.or) :{.0()±().2(j 2.:}(J±0.18 2.rv2±().19 N(;C 9S5 PI ().78±().(),'5 2.1K±0.17 i.()8±0.17 2.7(j±().17 Pi i.2r)±o.i() 1.92±().2() ().88±().07 1.90±(). 11 1^2 1.81±().ll i.r)i±().i2 0.i;}±0.0{i 2.():}±0.17 All) 10 PI 2.72±0.1H •i.8()±().:io o.ioio.or) 2.2()±0.2r) PI 0.79±().0K 3.8(j±0.2;i ().(i9±().07 1.72±0.1(i 11 Zw 28 pr 1 .r)7±0.2S ;{.27±().(jr) 2.I5±().I9 0.91±0.20 IT 2.2(j±().l 1 a.28±().59 1.9r)±o.:«) I.19±0.2-l pr a. i {±0.02 2.78±()..''),'-) 1.78±().;j,'j 0.»:{±0.18 Pl- 2.i:i±().r2 :{.l0±0.(i2 2.90±0.r)7 0.79±0.17 1! 11/ i P1 i.ioio.ir) ;<.()9±0.15 {).;«)±().oi :{.09±0.1() P2 ().2()±o.{);} •l.7()±0.(j(i 1.85±().28 2.;{r)±o. 10 \Vi\ 1 PI ().13±().();{ :{.2()±().2(i 2.81 ±0.15 i.98±0.ir) "Ahii'iiskh imlicaic iliai (lie iniiii Kiilfcrs from Haliiicr alisurptiuii. A.s a iim.M'iiiii-iui' (In: t-rriiih in iliv il/i flux, and (lie .sul)s('(|u('ii( flux ratios, arc IukIkt iliaii in (lie spectra nnalfccicd l>y Bainicr al)sorplion. 'in units of lO"''" crg.s .s~' cm"'' Table 2.'2 (C'oiitimircl) /((0//],\.t-27) l{[oi 11]\ Aisir)i>,r)UU7) /([A'//].\.\(>r).18,ti581J Namo Slit Position" 1(11/:/)'' /(///?) /(///J) HUH) VII Zw 100 PI ().(j2±().()l i.9i±o.ir) 2.11 ±0.13 O.G(j±O.Oi PI 1 .'iaiO.OH 1.75±o.or) i.8r)±o.os 0.70±0.0r) P2 ().SK±0.()5 a. U ±0.22 2.r>9±o.ir) 0.72±0.0r) A boll 70 PI 2.01 ±0.15 2.99±0.29 1.99±0.20 1.-I7±0.2a PI 2.23±0.;J3 3.11 ±0.32 1.09±0.1« 1.11±0.1S P2 ;{.9(j±o.ir) 2.il±O.IO l..')l±0.2.'j 0.91±0.17 p;3 r).;{K±0.r)0 2.70±0.;{.') 1.79±0.2S 1.:J2±0.21 Pl- i.7r)±().i:i :i.20±0.r)0 2.:»)±0.:J9 I.:M±O.2S "Aslcrisk.s iiulicalc (liat tlic hpci lriiMi suircis fioiii Ualmcr al)sor|)lioii. As a consciiiu'iicc! (In; errors in ilie 11/i Ilux, and the sul)s<'i|\ieii( (lux ralUis, an- lii^lu-r than in the spectra tiiiaifecteil !>>' Ualmcr alisurptiun. ''ill iiiiils ol" lO"''' ergs s~' cm"'' 'lal)l(' 2.3 l)(M(>(l(liMU'cl Line Strciigtlis /((o///l.\.\i!jr>u,,••>1)1)7) /([A'//].\.\tir>i8,(iri8i) Niiiiic Slil i'osilioii" Av (mag) l(ilJ)'' i:v 11 IM 0.21 l.KHiO.ll :i.71±0.2() 2.58±0.23 1.57±0.10 PI 1.09 2.21 ±0.11 7.28±0.57 2.2()±0.25 1.38±0.13 i'2 0.15 2.09±0.15 •i.;}9±o.;};{ 1.0()±0.10 1.15±0.09 P2 0.S2 2.80±0.20 5.01 ±0.39 2.25±0.12 1.37±0.10 \\i o.o;{ 0.I8±0.07 2.72±0.21 l.93±0.21 1.99±0.30 PI 0.17 1.20±0.12 2.51 ±0.29 2.21 ±0.25 0.95±0.12 Pf) 0.71 1.2;i±0.11 •t.71±0.:51 2.00±0.13 1.05±0.10 P5 1.17 1.78±0.18 •1.1G±0.3() 2.2G±0.17 l.G7±0.13 NCC 9sr) Pi 2.07 7.50±0.18 ;5.Ks±o.;«) I.5G±0.10 1.33±0.08 Pi i.r)i (j.79±0.51 2.9;}±o.;n 0.81±0.07 1.12±0.0G P2 2.17 27.0±1.7 ;{.00±0.22 0.38±0.05 I.10±0.07 Ar|) 10 1.52 2.91±0.19 7.11±0.1() 0.1G±0.05 1.32±0.09 PI O.SI 1.98±0.20 •I.SH±0.29 0.G7±0.07 1.28±0.12 11 Z\v -iS Pl- 0.S5 ;}.97±0.70 •i.i:}±0.80 2.37±0.17 0.70±0.15 pr l.OS 7.:i7±i.:« •i.i:{±0.K0 1.87±0.31 0.81 ±0.10 IM- 0.19 5.;J7±1.0(j 3.19±0.(i3 l.75±0.31 0.78±0.15 Pl- 0.77 5.(il±0.97 3..S.|±0.7(i 2.81 ±0.55 0.G0±0.13 1! Ih. 1 P1 1 .;{8 5.00±0.(i8 1.51 ±0.07 1.80±0.21 1.90±0.28 P2 1.00 0.00±0.09 0.21 ±0.87 1.77±0.2G 1 .«5±0.2 1 \VN 1 PI 1.25 l.()S±0.12 1.01 ±0.37 2.08±0.1l 1.27±0.I0 "Aftlcrisks iiidicalc llial (lu- hpcclruni siiircrs fViini iialiiuT ahsorpiion. Aa a ('oiisciiut'iict- llic criorn in (lie 11,^ lliix, and tlic .siil>M-(|nrnl Ilux ratios, arc liit^licr than in I lie .ipcctra iinalfcctcd liy Maimer ah.sorplion. ''in iinil.s of ergs s"' cm"'' 'lal)li' 2.;} (CoiitimuHl) /(lA'i/).\.\tir)18,(i58l) Naiiio Slit Position" Av (iiiHK) 1(11/^)'' nihi) l(llii) VII Zw K)() PI 1.10 2.0()±0.i;} 2.():}±U.20 2.0:{±0.13 o.-i^io-oa PI 0.22 I.MIO.IO 1.8(j±o.or) l.83±0.08 0.70±0.0r) P2 0.21 l.ll±0.0(j ;i.(ji±o.2;{ 2.r)7±o.ir) 0.(j()±0.05 Alu'll 7(j PI 2.10 20.1 ±1.5 r).3G±o.r)2 1.8;{±0.18 0.70±0.11 PI 1.1)9 7.;5r)±i.O}) •1.21 ±0.13 l.OliO.lf) 0.7r)±0.l2 P-2 1.2.') 15.(i±1.7 1.17±0.21 0.(i0±0.ll [Ki 2.01 •is.5±.i.r) •1.72±0.()1 1 .()()±0.2() 0.().')±0.10 Pl- 2.(M l().2±l.l 5.()3±0..SS 2.I8±0.;{() o.(ir)±o.ii "AMt-i'iskh iiuliciiUr tliikt ilu' sptfclruin siitfcrs from Daliiicr alisorptiuii. As a coiibi'iiuciicf l)io errors in lliir ll/i lliix, ami llic .- 2-4.2 Empirical and Ionic abundances of Oxygen and Nitrogen The temperature sensitive line [OIII] A4363 is very weak and was not detected in our spectra. Therefore we have calculated the o.xygen abundances using empirical methods. Several investigators have developed their own calibration of the R23 inde.K = ([Oil] A3727 + [OIII] AA-19.59..5007)/H.:^) with o.>:ygen abundance (e.g. Dopita ic Evans 1986: Edmunds i; Pagel 1984; McCall. Rybski. Shields 198.5). In the mean, the three methods may be approximated by 12 + log(0/H) = 9.26.5 - 0..33.X - 0.202.\* - 0.207.k^ - 0..33."i.x' where x = log(R2.j) (Zaritsky. Konnicutt. Huchra 1994. hereafter ZKH). and we have used this approximation to obtain oxygen abundances. Uncertainties in the o.xygen abundances were estimated in two different ways. First, by propagating the uncertainties in the line strengths (as given in Table 2.3) through this polynomial in order to obtain a "formal" error. Second, by computing theo.xygen abundances from each of the three methods separately, and computing the rms deviations about the mean. We found that in all cases these deviations in the o.xygen abundances were significantly larger than the "formal" error. Therefore, as with previous authors (e.g. ZKH). we adopted these deviations about the mean of the three methods as the uncertainty in the derived abundances. We have calculated the ionic abundances of o.xygen to serve as a consistency check on the empirical results. The nebular temperatures were first estimated from the empirical calibration of R23 with electron temperature as determined by Pagel et al. (1979). This was necessary in order to restrict the range of possibfe abundances derived by the ionic method. Considering the paucity of empirical calibrations to determine nitrogen abundances, we have computed the ionic abundances of nitrogen also using the same empirically derived temperature. .-V good correspondence between our ionic and previously discussed empirical o.xygen abundances wiis taken as a sign that the temperatures we have assumed are reasonable and that the nitrogen abundances are reliable. Electron densities were calculated using the [SII] doublet line strengtiis. and in all cases were found to be consistent with the low density limit such that colli.sional de-e.xcitation is unimportant (Osterbrock 1989) for the lines of interest ([Oil] A;{727. [OIII] AA4959.5007. and [.MI] AA6.548.6.584). The electron density used throughout our computations is 100 ± 30 cm"'. The [OMC routine in IRAF was employed to calculate the abundance of an ionic species, relative to singly ionized hydrogen, using: •V(-V^') ^ /(A) (2.3) .V(//+) j[\) • where .\ is the abundance of an ion, X"^': I(A)/I(HJ) is the observed, reddening corrected line ratio: and j is the emissivity of the line. We assumed that the hydrogen is completely ionized such that = H. Total oxygen abundances were computed assuming that N'(0)/N'(H) = N'(On)/N'(H) + .\(Oin)/.\(H). In order to compute the total nitrogen abundances, we corrected for the pres ence of .N'lII by scaling up the computed N'll flux using N'lII = (Oni/OII)NII. Therefore, we adopted: •V(-V) \{on -f- onI) \-{sn) (2.4) -V(//) S[oii) s{H) • This is a reasonable because NTI and Oil have similar ionization potentials and .M and OI have ionization potentials that are almost identical with HI (but see also the cautionary comments in Garnett 1990). The corrections applied to the nitrogen abundances were in the range I8-6S'^ (a 4-1'^ increase over the abundance due to XII alone was typical). Table 2.4 contains the o.vygen and nitrogen abundances derived for the ring galaxies. Uncertainties in the ionic abundances were estimated by propagating errors in the line strengths through the calculations. 2.4.3 Photometry of Star Forming Knots The IR.AF routine PHOT was used to extract broad-band BVR and .IHK magnitudes for individual star forming knots. The photometry for each knot was performed by e.xtracting the flux from circular software apertures (see Figure 2.2) chosen to include most of the flux from each knot and the dimensions of these apertures is given in Table 2.-5. Local sky values were used otitside the ring. This means that the magnitudes for the knots includes light from both the knot and any underlying smooth continuum that might not be associated with the knot. :}9 Table 2.4 Oxygen Abundances and Oxygen-to-.\itrogen Ratios Name Slit Position Empirical Ionic Ionic log(.\VO) 12 + log(0/H) 12 + log(0/H) 12 + log(.\7H) LT 41 PI 8.62±0.ri 8.57±0.15 7.67±0.14 -0.91±0.15 PI S.25±0.10 8.27±0.1l 7.23±0.12 -1.04±0.14 P2 8.66±G.12 8.6l±0.r2 7.49±G.r2 -L.l2±0.r2 P2 8.4o±0.r2 8.4o±0.10 7.41±0.13 -i.04±0.i;} P:} 8.82±0.1o 8.69±0.16 7.92±0.14 -0.77±0.17 P4 8.S1±0.14 8.67±0.09 7.62±0.14 -1.05±0.15 Po 8.o7±0.10 8.o.5±0.11 7.40±0.15 -l.lo±0.14 Po 8.62±0.1l 8.58 ±0.10 7.66±0.14 -0.92±0.l;} NGC OSo PI 8.73±0.14 8.64±0.17 7.61±0.16 -i.o;{±o.ifi PI 8.9:}±0.12 8.74±0.16 7.65±0.1C -1.09±0.17 P2 8.97±0.ll 8.8:}±0.19 7.67±0.18 -1.16±0.14 Arp 10 PI 8.44±0.r2 8.48±0.14 7.;}0±0.15 -1.18±0.17 PI 8.72±0.14 8.66±0.14 7.5l±0.l:J -1.I5±0.15 II Zvv 28 PI 8.61±0.12 8.57±0.1l 7.28±0.i;} -1.29±0.12 PI 8.62±0.12 8.59±0.14 7.;};i±o.i4 -1.26±0.15 PI 8.79±0.14 8.f59±0.09 7.46±0.1l -1.23±0.10 PI 8.o8±0.11 8.54±0.08 7.23±0.10 -I.:}1±0.15 II Hz 4 PI 8.62±0.ll 8.59±0.17 7.69±G.ll -0.90±0.12 P2 8.4:3±o.r2 8.45±0.17 7.45±0.r2 -1.00±0.11 WX 1 Pi 8.ol±0.l0 8.49±0.r2 7.47±0.13 -1.02±0.09 VII Z\v 4fj6 PI 8.8:3±0.14 8.68±0.18 7.28±0.1Cj -l.40±0.17 PI 8.94±0.12 8.68±0.14 7.57±0.17 -I.ll±0.l5 P2 8.64±0.r2 8.58±0.16 7.;}1 ±0.15 -1.27±0.UJ Abell 76 PI 8.52±0.09 8.ol±0.08 7.16±0.14 -l.;35±0.09 PI 8.75±0.14 8.67±0.r2 7.;M±O.I;J -i.3;i±o.i;} P2 8.79±0.14 8.69±0.09 7.:i2±0.1o -i.;}7±o.r2 P3 8.62±0.ll 8.74±0.08 7.29±0.11 -1.45±0.1I PI 8.45±0.r2 8.46±0.09 7.09±0.r2 -1.37±0.12 Tabli' '2Ji I'lKoircitcd Culois uf Imliviihial Star Koniiing Knots Name knot Apor Si/c H-\ V-H li - K H-K V-K .1-11 II - K (aiTsi'c^) (mag) (mag) (mag) (mag) (mag) (mag) (mag) i;r 11 1 13 0.5.1±0.17 0..')7±0.13 l.ll±0.17 2.15±0.12 3.02±0.13 0.7()±0.11 0..')(j±0.11 2 i;{ 0.1(i±{).19 0.3<)±0.l() 0.K.')±0.19 2.19±0.1,'3 2..'j8±0.1.'') 0.-13±0.11 0.()1±0.I5 3 la ().r)()±().20 0.1(j±0.1(j 0.9()±0.19 2.G0±0.1I 3.0(i±0.1.') 0.9(j±0.l2 O.I9±O.I3 avc ().r)0±0.is 0.I7±0.1.') 0.97±0.S1 2.ll±0.ll 2.8S±0.11 0.71 ±0.12 0.r).'j±0.13 NCIC {)Kr> 1 2:J ().72±0.21 0.I7±0.13 0.1.19±0.23 1.91 ±0.13 2.3S±0.11 0.(j0±0.1l 0.I0±0.1-1 2 IH ().58±0.3l 0.12±0.1S 1.00±0.30 l..')9±0.15 2.01 ±0.17 0.70±0.1l 0.()0±0.11 avo 0.()5±0.1l 0.10±0.17 1.10±0.27 1.7f)±0.ll 2.19±0.1() 0.{).')±0.11 0.r)0±0.ii Ai|) 10 1 18 ().H2±0.()!) 0.27±0.0(j 1.09±0.0S ... _ 2 IH 0.7r)±0.12 0.2.')±0.09 l.00±0.12 a IS 0.7()±().12 o.23±o.oy 0.93±0.12 •1 2;} 0.5()±().07 0.12±0.0(i 0.(i2±0.07 T) 15 ().17±().13 0.23±0.11 0.70±0.12 () 23 ().57±().l 1 0..3(i±0.11 0.93±0.13 avo ().()3±().ll 0.21 ±0.0!) 0.HK±0.I1 11 Zw 28 1 2;} ().3{)±().()1 0.1{)±0.03 0..')S±0.01 2.10±0.01 2.3r)±0.01 0.17±0.03 0.32±0.01 11 llz 1 1 i:} 0.r)3±0.1!) 0.3S±0.15 0.91 ±0.19 2.12±0.13 2.80±0.11 0.78±0.11 2 13 ().r)7±().l3 0.3!J±0.10 0.9()±0.13 2..')K±0.09 2.97±0.09 0.01 ±0,08 13 o.rjsio.i!) 0.3(j±0.1.'") 0.91±0.17 2.12±0.13 2.78±0.ll 0..')2±0.13 avc ().5(j±().17 0.37±0.13 0.9I±0.17 2.I7±0.12 2.«S±0.r2 0.(i3±0.1l \V.\' 1 1 13 ().r)2±().i7 0..')9±0.13 1.1 l±O.I() 2 13 0.11 ±0.11 0.33±0.09 0.71±0.11 2.17±0.10 2.S0±0.11 0.7S±0.I1 0.22±0.l 1 13 0.I0±0.1 1 0..51±0.11 1.00±0.I3 2.18±0.11 3.02±0.ll 0.03±0.1.') 0.23±0.1.'') avc O.IOiO.ll 0. l.S±0.11 0.9.'')±0.13 2.IS±0.11 2.91 ±0.13 0.70±0.1.') 0.22±0.1l 41 f ^ iC rc IC 32 S •—• ci _ • sC Z; Z: -H i 5 5 +i 5 - —* r- 5 r- r- — === • vT^ *?4 —• o T X X c: i^C tiT LC —• w w w '3 o o ^5 d d w o , rS jH -H -H -H -H -H -H -H -H — :C Cl X X tC X i3 iC W tc iC uf »JC LC ^-* •* iO r— CM CM rt CM c: CM —M —* —• —• 'tc r"* d d o d d d d 1 4^ -H -H -H -H -H -H -H -H -H o t- r- iC r^ •M iC w C3 C5 C5 X; r- X T^ — CM* — — — — —' —• LC r- *T c; C; — — — —• o —• "2 O d d d d o , iw -H -H jH +1 +1 -H -h -H -H -H oc LC C5 T — TM — -—' t-- o r- w "C* — — — — — — —• — —N o CN . L.t' X X r- r? CM —M ^5 CM •>J o o d w o d d Nwi* o -H jj -H -H -H -H -H -H -H -f< -H -h -h -H o t- CM w r^ t- o -r o t?5 X C5 X O o 30 X , r- r- i-'C «. CM -- •w —• o — 'S —: CP d d d d 1 -h -h jj +1 -h -h -h -h -h -h -h -h -H -h T X r- c^ X X >3^ CM -v^ —7 •••^ — CM — •~r ?c o == o o d d ^5 d d C: C5 —. 7C L.C X X ;c __ — o —. — — — — — o o — — CM CM CM CM O d d d d d d d d d :z HH -h +< -h -h -h -h -h -h -h -h -fi -h -h -h ~ ~" uC t- C3 «ys w5 CM CM CM —' •-3/ 'M tc o cc CM CM T Lf ic O O d d d d d N; ir oc iO lC lO X tc tc 'A •• — -x , •rg rc uT t- X o _ >• -2 3 r^ w s: — — 5c < Table 2.0 Kxliiictioii ('diiccIchI Colors of liuliv'ulual Star l-briniiig Knots N a lilt' knot Av" \i - V V - K H - K u - k \' - k j - ii ii - k (mag) (mag) ("'ilS) (mag) (mag) (mag) (mag) i;r 11 1 0.21 0.17±0.17 0.52±0.l;{ 0.99±0.17 2.32±0.12 2.81±0.13 0.7I±0.1I 0.55±0.11 2 o.ir) 0.;12±0.19 0.2K±0.10 0.59±0.19 1.91 ±0.15 2.18±0.15 0.39±0.11 0.58±0.I5 3 0.K2 0.21±0.20 0.20±0.10 0.19±0.19 2.08±0.11 2.;M±0.15 0.89±0.12 0.i;{±0.13 Ncc ysf) 1 2.17 -0.07±0.2l -o.mio.kj -0.21 ±0.23 0.35±0.i;} 0.20±0.1l 0.;}7±0.1l 0.22±0.1l Arp 10 1 O.K-l 0.2:i±0.07 -0.09±0.00 0.11±0.07 0 1.52 0.08±0.11 -0.02±0.11 0.0()±0.i;J 11 Z\v 2S 1 1.15 0.02±0.0I -o.io±o.o;{ -O.0K±0.01 1.1i±0.0i i.:m±o.oi 0.37±0.03 0.21±0.0l II 11/ 1 r 1M 0.08±0.19 0.03±0.15 0.12±0.I9 1.55±0.13 1.58±0.1l 0.08±0.1 1 2- i.;is 0.12±0.13 O.OliO.lO o.i7±o.i;} 1.71 ±0.09 1.75±0.09 0.51 ±0.08 :r i.:{s 0.l:i±0.19 0.01 ±0.15 0.15±0.17 1.55±0.13 1.50±0.11 0.12±0.13 \VN 1 1.25 O.OOiO.M 0.22±0.ll 0.28±0.i;{ 1.09±0.11 1.92±0.1l 0.51 ±0.15 0.11 ±0.15 VII Zw •!«() 2- 0.21 0.22±0.10 0.27±0.0S 0.19±0.10 1.52±0.15 1.79±0.15 0.50±0.10 0.29±0.10 ;r 0.21 o.2(j±o.i;} 0.29±0.11 0.55±0.12 5 1.10 o.;m±o.i;{ 0.17±0.10 0.50±0.i;{ 1.0I±0.17 1.20±0.17 0.59±0.08 -0.10±0.17 H- 0.22 o.ir)±o.U8 O.l l±0.07 0.2S±0.08 0.95±0.13 1.70±0.12 0.50±0.08 0.19±0.13 9' 0.22 0.17±0.08 0.17±0.07 o.:{;i±o.OH i.oo±o.i:i I.78±0.I2 Al)('ll 7(j 1 2.01 -0.17±0.2(j -0.08±0.1 1 -0.27±0.2;{ 0.27±0.10 0.19±0.12 0.37±0.05 0.03±0.00 2 2.o;{ -0.12±0.21 -0.12±0.12 -0.25±0.2I 0.11±0.09 0.01 ±0.10 0.31)±0.05 0.00±0.0() 1.25 0.11 ±0.20 0.01±0.1 1 o.is±o.2;j 0.0I±0.10 0.09±0.12 0.I2±0.05 0.10±0.00 "A.stciisks iiulicatc thai llu- Av is an avi-ram' foi { losfly spacrtl knui.s However, as we shall describe later, our analysis of the color evolution of the knots takes into account any possible underlying old stellar component which might be associated with a stellar density wave in the underlying disk. Our photometric errors were estimated in the usual way. taking into account photon counting statistics from source and sky. as well as uncertainties in the flat fielding, which was more significant in the IR observations. From the magnitudes we calculated B - V. \' - R. B - R. R - K. V" - K. .J - H. and H - F\ colors. Table 2.5 contains the undereddened colors of the individual star forming knots. Corrections of the colors due to internal reddening were applied only in those cases where there was overlap of the spectra with the knots, such that a reliable .A.v could be calculated from the observed Balnier decrement. Table 2.6 contains the dereddened colors of that subset of corrected knots, along with the Av used to correct for e.Ktinction. [u the discussion of knot ages, we refer only to those knots with reliably determined .-Vv s. 2.4.4 Oxygen and Nitrogen Abundances We find that our sample of ring galaxies have sub-solar metallicities in the range one-half to one-fifth solar in [-V/H] and [0/H] ratios. There is a suggestion of an increase in the mean nitrogen abundance for rings of larger linear size, as shown in Figure 2.3. The oxygen shows no trend with ring diameter, and is approximately one-half solar for the entire sample (see Figure 2.1). In this scction wc discuss these results, first by ignoring the effects of chemical evolution in the ring waves, and then by exploring the possibility of in situ production of oxygen and nitrogen in the rings. Are the observed abundances "normal." ignoring the possibility of chemical enrichment in the ring wave? Figure 2.5 contains a plot of observed (ZKH: \'ila-Costas 1992) o.Kygen abundance gradients in a few representative disk gala.xies. covering a range of Hubble-types. The oxygen content of the rings is consistent with them lying atop gradients in progenitors of differing Hubble-type. Interestingly, there appears to be a morphological segregation of Hubble-type with ring size. The smaller rings inhabit small late-type gala.xies (T-typcs (J-X) whereas the large rings inhabit larger mid- to early-type galaxies (T-type 2-5). 44 We note that this result is probably due to a selection effect. Large rings can t e.xist in small gala.xies and small ring waves in large galaxies may be misidentified as tightly wrapped spiral arms or inner, hot-spot rings. In Figure 2.6 we show plots of o.xygen abundance versus the nitrogen-to-o.Kygen ratio with and without error bars, respectively. Interestingly, the three smallest rings have the smallest .\/0 ratios whereas the larger rings have larger .\/0 ratios. The increase of .\"/0 (and thus the increase of nitrogen) with ring diameter is consistent with the segregation of Hubble-type with ring size, since late-type galaxies have lower .\/0 ratios (on average) than mid- to early-type gala.xies (V'ila-Costas Edmunds 1992). What is the role, if any. of chemical enrichment in the ring wave? Vila-Costas Edmunds (199.'J) suggested that the time-scale for the release of freshly synthesized nitrogen is probably > 5 X 10^ years. The dynamical ages of the rings are typically 100 to ;}00 Myr (150 Myr is approximately the average). Therefore, it is probable that the nitrogen was produced in pre existing stars before the collisions that formed the rings. However, the question of the origin of nitrogen (see Figure 2.6) in these systems is complicated by the fact that o.xygen may be produced in the rings (see below). Without knowing (accurately) the oxygen abundances of the progenitors, before the collisions, it is difficult to conclude whether the production mechanism of nitrogen in ring gala.xies is mainly primary, secondary, or a mixture of the two. Previous Ha imaging (NfA) of the sample has revealed that the total star formation rates (SFFls) in the rings are independent of ring diameter, suggesting that the SFRs remain constant in the ring waves as they propagate into the disks of the targets. If the SFRs remain constant, what is the expected mass of o.xygen produced in a 150 .\Iyr year period (this time-scale was chosen because it is a typical dynamical age of the rings in the sample)? Oxygen is produced primarily in massive stars (> 10 .\1^,) which live on time-scales of a few x 10' years and explode as supernova. In the following calculation, a Salpeter IMF was adopted and normalized to a constant SFR of 4.1 .\I^. yr~^ (the average global SFR (M.-V) of the sample). We assumed closed-box chemical evolution and that the o.xygen released in the supernova is instantaneously nii.xed throughout the disk of the ring gala.xy. The mass of oxygen produced 1 1 1 1 1 1 1 1 1 1 1 1 1 1 j 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 8.5 •WN1 A||Zw28 ~ •Vll Zw 466 DAbell 76 - OArpIO •LT41 •II Hz 4 TirNGC 985 — solar 1 I half solar k i 1 1 : "J 1 7 r - ( — I 1 ; ;• quarter solar "j 1 o* - !f :: i I ( i i i f I J I [ ] 7 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 5 10 15 20 25 30 35 Major Axis (Kpc) l''i}i,iirc 2.:i A plot of tlic nitrogen abnnilancc.s against the k'liglii of the major axis, 'i'lic ahnndanic orcacii star forming iinol is rt'|)rcs(«nt('(l by a gcoiiH't rical symbol and a key is provitlcd to icifntily tlii' ring galaxies. Diameters were calcnialecl a.ssuining II,) = 75 km s~' Mpc"'. 1 I I I I I I I I I I I I I I I I I I I I I I I I I I I I I \ I r 9.5 I ' ' ' ' I •WN 1 AIIZw 28 •VII Zw 466 •Abell 76 OArp 10 •LT41 •II Hz 4 •NGC 985 9 •jr ROlar. .5r. I g "5 if- ir o + .h9ll solar.. CM ^ 8.5 I O cff .qwarter.5piar. 8 J I I \ I I I I 1_J I I I L_J I L J I I L_L 0 5 10 15 20 25 30 35 Major Axis (Kpc) I'iSiirc 2.1 Similar to Figuri' 2.;}, oxcopt that in this ligiirc tlu' oxyg^'ii al)iiiitlan(i' is |)k)tl('(l against the length ol' the major axis. I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I ^ TT 9.5 • WN 1 A II Zw 28 • VII Zw 466 • Abell 76 O Arp 10 "1141 • II Hz 4 * NGC 985 Tr I g. CT I- 8 — 7 6 -- 5 - - 4 ... 3 -- 2 --- I I I I I I 1 1 I. I I I I I I I I I I I I I I I I I I I I I 10 12 14 16 18 20 Radius (kpc) •igiirc 2..') Siiniliir to l-'igiirc 2.1, c.xci'i)! that the o.xygcii abiindaurc is plolti-d against the radius of the major a.sis. Ovcriayt'tl arc observed abundance gradients, lor a range of liul)i)ie-ty|)es. lOacli T-lype is repre.seiited by a didereiit line, and a key is provided to tell wliirli line belongs to what llnbble-lype. Data for these gradients was taken from ZKII and \'ila-('ostas iV l-xlmuiids (l!)<)2). Note llie segregation between llie llul)l)le-ty|)es of the small rings and the larger rings (.see te.xt). 48 is obtained by multiplying 150 Myr by the integral of the convolution of the normalized IMF. «SFR '^(m). with the yield (in solar masses) of oxygen as a function of stellar mass, y(ni) (y(m) from Woosley Weaver 1995): rlOO -^^oxygen = 1-5 X lO'^ / osFR ^("0 !j{ni) dm = 6.9 X 10'' Mi. {•-^-5) Jw In order to compute the o.xygen abundances relative to hydrogen we needed to make a reason able estimate of the mass of hydrogen contained in ring galaxies. The average global HI mass in ring galaxies is. approximately. 5 x 10^ .\I .- (this average was calculated from single dish and interferometric HI observations in .\ppleton et al. 1996: Charmandaris .\ppleton 1996: Horellou et al. 1995; .Jeske 1986). CO observations of rings reveal that large amounts of H^ can exist in ring galaxies (Higdon. Rand, iij Lord 1997: Gao et al. 1997: Horellou et al. 1995). Observed Hj masses are on the order of a few x lO'^ M..-,. and the average H^ mass is about 5 X 10^ M . (e.g. Horellou et al. 1995). Thus, the total average hydrogen mass of a ring gala.xy is generally between 5 to 10 x 10^ \[ ;. We use this range to calculate the amount of oxygen relative to hydrogen. The lower value results in: 6 0 X 10'' \/ 12 + log{0/H) = 12 + log ^ ox IQ-' .V/. which is 3 rimes the observed mean o.xygen abundance for the entire sample (S.G5). The upper value results in: 6 9 X lO'' \I 12 + log{0/H) = 12 + log ^ = 8.8.3. (2.7) which is approximately equal to. or less than, the o.xygen abundances of a few of the knots in some of the ring gala.xies. Flowever. an o.xygen abundance of 8.83 remains 1.5 times the observed mean. In spite of the uncertainties inherent in this calculation, we conclude that it is possible for oxygon to be produced in situ in the ring waves. The fact that the observed o.xygen abundances are slightly (1 to 3 times) lower than predicted by our simple calculation may result from oxygen being produced and. i) blown out of the disk, to rain back down onto the I I I I I I I I I I I I I I I I I > I I I I I I I I I I I I I I I I I • WN1 A |IZw28 • VII Zw 466 • Abell76 O ArpIO • LT41 • II Hz 4 tV NGC985 -.5 — -1 — O -1.5 Pure Primary -2 J ^ I I I ^ I I I I I I I I I I I I I 7.6 7.8 8.2 8.4 8.6 8.8 9.2 9.4 12 + log(0/H) i-'igiirc 2.(i Plots ofliic o.xygcii alMiiKlaiicc versus llic iiilrogcii-lo-o.\ygcii ratios, witli and willioul tlic errors, respeclively. 'I'lie (la.slied lines indicates the areas where nitrogen prodnclion is |)iirely primary or jxirely secondary, and the solid line is the sum of the two. 0 1 I I I I 1 I I I I I I I i I I I I I I \ I I—I—I—I—I—I—I—1—I—I—r -• WN 1 A ||Zw28 VII Zw 466 • Abell76 O Arp 10 • LT41 # II Hz 4 ic NGC 985 -.5 — O z -1 — o O -1.5 - Pure Primary J I L d 1 L_J 1 L_J I I I I _L -2 I I L J I 1 1 I I I 1 I I I 7.6 7.8 8 8.2 8.4 8.6 8.8 9.2 9.4 12 + log(0/H) I'if^uic 2.(j (('oiiliiiiK'd) 51 disk some time later (i.e.. rings aren't closed boxes), or ii) hidden in a hot. unobservable lO*' K phase (e.g. Kolbunicky Skillman 1997). Large azimuthal abundance variations are neither e.Kpected nor observed in disk gala.xies (e.g. Martin Belley 1996), thus another clue that there might be in situ production of oxygen (and nitrogen!?) is the large observed spread in the abundances from knot to knot around some of the rings (particularly in .\bell 76 and LT 41. see Figures 2.3 and 2.4). However, considering the uncertainties (±0.3 dex on average) of empirically derived abundances, we cannot confirm the reality of the variations. Higher signal-to-noise spectra, capable of detecting the weak temperature sensitive [OIH] A4363 line, are needed in order to determine if these abundance variations truly e.xist. Finally, we wish to point out that it is possible that not all the rings formed in spiral galaxies. H Zw 28 has been shown to have a fairly quiescent. HI disk and may have been a barred .Magellanic irregular gala.xy before the collision (Appleton et al. 1998). .-Vbell 76 ha^i a mass inside the ring of 1.9 x lO'^ M r. although it is likely that the ring lies in the inner regions of a larger disk (.Marston k. Bransford. in prep). 2.4.5 Evidence for two bursts of star formation in II Zw 28 and Abell 76 II Zw 2S shows significant Balmer absorption in its spectra indicating the existence of a substantial A-star population. .V representative spectrum of II Zw 28 is shown in Figure 2.2. The existence of strong narrow hydrogen emission lines superimposed on broader absorption features, coupled with the blue broad-band colors of the knots (see §2.2). shows that II Zw 28 is continuing to form stars. HST observations (.\ppleton et al. 1997) show that the current star formation is both in a narrow ring and within a bar-like structure just inside the ring. .Vote that the emission lines are not symmetrically superimposed on the absorption lines. In particular, the HJ and H-. emission lines are blue-ward of the center of H.i and H* in absorption, suggesting that the .\-star population and the bursting knots are not coupled kinematically. In some ways II Zw 28 resembles a .Magellanic irregular gala.xy and if .so. it it likely that the .-V-spectrum originates from the bar-like structure and not from the ring. Only 52 one slit position was obtained for this galaxy and so further work will be necessary to confirm this possibility. Abell 76 shows significant Balmer absorption in knot 2. speculated to be part of the com panion involved in the collision seen in projection against the ring (Talent et al. 1982). Talent et al. speculated that Balmer absorption might be affecting their spectra, without actually observing it. We confirm their speculation, in that knot 2 contains a substantial .-V-star popula tion. .\s in n Zw 28. there are narrow hydrogen emission lines superimposed on the absorption features suggesting on-going star formation. 2.5 Ring Kinematics [n earlier papers we have presented the kinematic data for two of the rings. .-Vrp 10 (Char- mandaris .-Vppleton 1996) and V'H Zw 466 (.-Vppleton. Charmandaris remaining galaxies observed, only two. LT 41 and .\bell 76. had sufficient coverage of bright knots to allow the kinematics of the ring to be explored, .\bell 76 will be discussed in a separate paper and so we present here data for LT 41. Of the others, in most cases only two or at the most three knots in the ring were observed and this is not sufficient to allow us to model-fit the possible expansion and rotation of the ring. However, in Table 2.7 we present the velocities of those knot for future reference. We note that for II Zw 28. the velocity of the absorption lines is seen to vary .systematically across the bar-like structure. The absorption-line velocities are given as a function of spatial position along the bar in Table 2.7. 2.5.1 The Kinematics of the Ring in LT 41 In one .system. LT 41 (the highest redshift ring, z = 0.07) we were able to extract enough velocity measurements around the ring to allow for a fit to be made to the velocity field. In a collisional ring gala.KV, an outwardly propagating density wave is produced such that the gas (and stars) should be simultaneously expanding and rotating. In this section we verify that this simple model adequately fits the velocity data derived from the spectra of LT 41. The position angle of the major a.xis was found to be 120° (.\ thru E). We have made the simple assumption that the observed ellipticity of the ring is a result of observing a perfectly Table 2.7 Velocities of Star Forming Knots Xame knot V'el from: Hcv (em)" H.i (ab) H7 (ab) (km s~') (km s~') (km s~') .\GC 98.5 1 13.010 H Zw 28 I' 8.6:}0 8.8;}0 8.660 8.590 8.7.50 8.700 6" 8.570 8.680 8.600 9" 8..560 8.700 8..570 [[ Hz -t l-.-jf L2.810 VV.\ I :} 16.;580 ".•Vsterisk indicates that, in [I Zw 28, the velocity information is first given for the east side of knot I. and works west in 3" increments, f indicates that the velocity for these closely spaced knots, in II Hz -I. are the same. .-Vlso. (em) denotes velocities from emission lines, whereas (ab) denotes velocities from absorption lines. circular ring from an inclined viewing angle. From the published Ha and B-band images of LT -II (.\I.\: .\\I) we estimate the inclination of the ring to be i = 44 ± 5°. The FWHM centroids of the Hn emission line were measured azimuthally around the ring (0 degrees azimuth was chosen to be the optical major-a.xis of the ring). The errors in the velocities obtained wore ~40 km s"'. We have performed a least-square fit of the velocity centroids as a function of deprojected position angle around the ring, using; f(6') = V'syst + sin i (V^oiCosO + V^^p.'^inO) (2.8) where \'syst is the systemic velocity. V'mt the rotational velocity. the radial e.Kpansion velocity, and sin i is the sine of the inclination angle. Based on the fit. shown in Figure 2.7. we verify the collisional nature of this ring and that a simple model of an e.xpanding and rotating ring is appropriate. The fit gives a reduced of 1.52 and. V'syst = 22.190 ± 11 km s~'. \'r„t = .{.M ± 25 km s~'. and \'e.vp = i km s~'. .\n estimate of the age of the ring can be obtained from the expansion velocity of the ring, .\ssuming that the velocity of the ring remains constant as it propagates outward, the dynamical age (rayn) of the ring is: 22500 22400 *—1 I 22300 tn a M 22200 -fj o o i-H (U > 22100 22000 21900 0 50 100 150 200 250 300 350 Deprojected Position Angle (Degrees) I'igiirc 2.7 A U'lisl-Hciuarc lit to tlir l-'W'llM velocities of lln against the tieproJectecl a/iiniitlial position angle in I.T II (zero degrees a/iiiiiith is the optical iiiajor-axis). Tlie errors on the velocities used in the lit ar<' ~l() km s~', 55 R ^ where R is the linear radius (12.9 kpc) of the major a.Kis (assuming a distance of 296 Mpc). .•Vssuming that in the collision the main component of the ring perturbation is to lead to a radial impulse in velocity and that the azimuthal component can be decoupled (at lea^t to first-order), then we can use the azimuthal velocity to estimate the mass in the gala.Ky. For a strongly off-center collision this approximation breaks down, but the case of LT 41 is probably only mildly off-center, since the nucleus is only mildly displaced from the center of the ring. Making the above assumption, we obtain a value for the mass contained within the radius of the ring (rotating interior to the ring). Mt: V'" ft .\[j = ~ 6.5 ± 0.7 X lO'' (2.10) Ci where G is the gravitational constant. We note that the ratio of V'exp to V'rot in LT -t 1 is 0.4. typical of most ring gala.xies which span a range from 0.2 to 5 (see .-Vppleton Struck-.Marcell 1996). 2.6 Colors of Star forming Knots and Comparison with Numerical Models of Stellar Evolution 2.6.1 Description of Isochrone Synthesis Models Leitherer Sc Heckman and Bruzual Chariot have built isochrone synthesis models to follow the evolution of stellar populations (LH95: BC95). The models use theoretical (e.g.. .Maeder 1990: .Maeder .\Ieynet 198S). or observationally constrained (see BC95 and ref erences therein), evolutionary tracks in order to compute a star's position in a theoretical color-magnitude diagram (C.MD) as a function of the age of the burst. The magnitudes (i.e. B\'R and .JHK) of the cluster can then be computed at any arbitrary point in the evolution of the star cluster. V\e will use the result of such computations for comparison with our data. 56 There are several clifTerences between the LH9o and BC95 models. Most importantly, the LH9o models include contributions from nebular lines, and specifically free-free and bound- free emission from hydrogen and neutral helium, and the two-photon continuum of hydrogen (LH95). The BC95 models employ a pure stellar continuum, and therefore their models are generally bluer than the LH95 models. Because theoretically determined C.MDs for stars on the .\GB are uncertain. BC95 opted to use observationally constrained CMDs for those stages of stellar evolution. By adopting empirically determined CMDs for .-VCIB stars, their models could be evolved to ~20 Gyr. Leitherer and Heckman's approach to this theoretical problem was to stop evolving the models at ~300 .\Iyr. The LH9o models were developed to follow the synthetic properties of starburst systems and were optimized for stars with masses greater than 5 . and should not be used for older, more evolved stellar systems. Finally in BC95. only solar metallicity stellar evolutionary models were used, whereas the LH9o models cover a range of different metallicities: O.L. 0.25. l.O. and 2.0 Z;. We employ the instantaneous starburst models of LH95 with Z = 0.25 and 1.0 Z,. and the instantaneous burst models of BC95 for comparison with our data. To consider the case of a starburst occurring on an old star density wave (the original Lynds and Toomre (197(i) picture) we have modified the BC95 models results to include differing burst strengths. The burst strength is defined as the mass of new stars to the total mass of the composite system. Thus, the burst strength, b. can range from b = 1 (no old stellar population, pure burst population) to b = 0 (pure old stellar population, no burst population). We built models with the following: b = 0. b = 0.001 (i.e. 0.1% of the total mass is in new stars), b = O.Ol. b = O.l. and b = I. We made two sets of models: those with a 15 Gyr underlying population and those with a I Gyr underlying population. The L Gyr population was thought to be more appropriate for systems like II Zw 28 and .-Vbell 76. where Balmer absorption lines suggest a post-.starburst population. These models were then compared with the de-reddened colors of the star forming knots. The LH95 models were not evolved to sufficiently long times to allow us to modify their models in this way. Before discussing the results of this comparison it is worth spending some time discu.ssing •57 the potential effects of reddening on our results. .-Vlthough we have data for many knots in the gala.xies (see Table 2..5) we only compare our results for those knots where a reliable Balmer decrement is available. In column ;{ of Table 2.6 we give the calculated values of .-Vv based on the decrement. In a number of the gala.xies the values of .\v are small. However, in three cases the values of Av are large. The largest e.xample is the knot in the ring of XGC 9.S.5. This knot has an .-Vv of almost 2.-5 magnitudes. This is in agreement with the work of Rodriguez Espinoza Stanga (1990) who also obtain a high .\v for this gala.xy. Clearly this gala.vy is very dusty (it has the highest CO flu.x of the gala.xies observed by Horellou et al. 199-5). The other very large extinction is found in .\bell 76 (knots I 2) where the values of .-\.v are larger than 2. We note another interesting case, that of \'II Zw 466 (an early photometric study of this gala.xy was made by Thompson Sc Theys (1978)). It was suggested some years ago by Toomre (see .-Vppleton. Charmandaris and Struck 1996 for discussion) that the knot slightly interior to the ring might be the remnant of the nucleus of the original target galaxy since it is rodder than the other knots. Oddly the knot was observed to be very weak in our K-band image (Paper I) and this led us to suggest that the knot is in fact an H II region. Our spectra confirm this belief as can be seen from Table 2.6. The knot in question is knot -5 which has an unusually large Balmer decrement, relative to the rest of the knots, .\fter correcting for the reddening, it is seen that the color of the pos.sible nucleus is only slightly redder than the other knots in the ring at B-V. and is indistinguishable from the others as one moves into the infrared. We therefore believe that \TI Zw 466 is a truly "centrally smooth" ring (i.e. has no nucleus) and may bo tlie best example of its kind so far discovered. We al.so wish to comment on .Vbell 76 and II Zw 28. galaxies for which Av may be effected by the pre.scnce of Balmer absorption. We have discussed in an earlier section how we attempted to correct for this by fitting under the emission line. In order to provide some consistency chcck. we used the .\v derived from the Ho/HJ ratio to correct the H-//HJ ratio. The results were that, in II Zw 28. the extinction corrected H-/HJ ratio is 0-54 (0.47 is the theoretical value), in agreement at the Lo'^ level. In .\bell 76 the dereddened H7/HJ ratio is 0.61. in agreement with the theoretical value at the :iO% level. We have put an asterisk against those knots in Table 2.6 for which Balmer absorption may effect the results. In general the effect would be to cause us to overestimate the reddening, and therefore to over-correct the colors of the knots. 2.6.2 Results In Figures 2.8 through 2.10 we plot the models of LH95 and BCOo on color-color diagrams, using a Salpeter LVIF. overlaved with the reddening corrected colors of individual star forming knots in the rings. .-VIso in Figures 2.8-2.10 we have plotted the models without the data in order to make it easier to read the figures with data, and have labeled a few key time intervals. Comparison of the colors of the knots in the rings with the isochrone synthesis models revealed that they have ages that range from 4 to 80 .\Iyr. less than the dynamical ages of the rings (100 to 300 Myr). Interestingly, the knots from individual ring gala.^ies cluster around the same, very young age which argues that their formation is recently triggered by a coherent event all around the ring. This strongly supports the production of the knots as being due to the propagation of a radial density wave. The fact that the young star forming regions in an individual ring gala.xy are coeval, and of the same metallicity (within the errors), makes ring systems fascinating objects for studying stellar evolution. Further, the reddening corrected colors of the starburst knots in the ring gala.xies agree very well with both of the stellar evolution models. This result is very satisfying, in the sense that we have demonstrated the utility of using ring gala.xies as litmus tests of theoretical models of stellar evolution. We caution the reader that the knots may suffer from mild dust emission, in which case the .1 - H and H - K colors would be significantly affected. .\ mild dust component would be seen primarily in a reddening of the H - K colors (e.g. see Figure .{ of .-Vlonso-Herrero et al. I99(j) of the star forming knots. In the H - K versus J - H color-color diagram (Figure 2.10). the evolution occurs over a very small range in magnitudes, and can be affected by dust emission, which can limit the usefulness of these two colors as an age indicator. .Vevertheless. agreement between the models and the H - K and .1 - H colors of the knots is remarkable for most of the I I I I I I I I I I I I I I I I I I I [ I I I I I I I I I I I I I I I I r I I I I I I I Ih. Z = 0.25 Zq Ih. Z = Zp be, Z = Zo ^ 1 Myr • 4 Myr • 10 Myr • 100 Myr O) ra E m o -.5 -1111 I I I I 1 I I I 1111 1111 I I I I I I I I I I I I I I I I -.5 .5 1 1.5 2,5 3.5 V • K (mag) I'igiHT 2.K III this ligiirc, and in i-'igiiros 2.9 iV, 2.10, we plot tlii' models of IJI9r) and B('})5 on color-color diagrams without and with data. In the plots without data we lahel a few key time steps (1, I, 10, 100 Myr), which is done in order to ease the readinjj, of the |)li>ts with data. 'I'he models, ploltetl with ilata, are shown with time sle|)s ol 1, 2, I, (i, S, 10, 20, 10, ()0, SO, and 100 Myr, and are overlayed with the dereddeiied colors of individual knots in the rings. Mach plot with data has a geometrical .symbol to denote a time step in the evohition ol" the model: the live pointed stars for the solar metallicity IK'9.') model, three jjointed stars for the .solar iiietallicily 1-1195 model, and the plus sign for the one-quarter solar metallicity 1.119r) model. 'The models generally evolve from the lower left t(j the upper right. This lignre contains the \' - K versus M - \' color-color diagram. TH I I I I I I I I I I i I I I I I I I I I I I I I I I I I I I I I I I 1 h- • WN 1 A ||Zw28 • VII Zw 466 • Abell76 O Arp 10 • LT41 • II Hz 4 •Ar NGC985 Av = 1 mag I—^ -.5 I I I I I I I -1 -.5 0 1 1.5 2.5 3.5 V - K (mag) i'igiirf 2.S (("oiitiiiiu'd) 1 r 1 r "1 r 1 — Ih. Z 0.25 Zo Ih. Z be, Z ^ 1 Myr • 4 Myr • 10 Myr • 100 Myr .5 — O)ro £ > GQ 05 0 — -.5 J L 1 -I L -.2 .4 .6 V - R (mag) I'igurc 2.9 \' - H versus li - \' coloi-colur diaffianis. .Sec caption lor I'Mgiirc 2.S. t "I r ' ' I • WN 1 A ||Zw28 • VII Zw 466 • Abell76 O Arp 10 • LT41 • II Hz 4 •k NGC985 -T Ay = 1 mag J L J 1 L J L I I I -.2 0 .6 V - R (mag) "2.9 (('oiiliiHU'tl) Ih, Z = 0.25 Zo A 1 Myr Ih, Z = Z© • 4 Myr be, Z = Zo • 10 Myr • 100 Myr v; -.2 0 .2 .4 .6 .8 H - K (mag) 2.10 II - K versus .1 - II tulur-rvilor iliiifjjrfiiiis. Sec (•a|)l'u)n for i'if^urc 2.H. • WN 1 A II Zw 28 • VII Zw 466 • Abell76 O Arp 10 • LT41 .8 • II Hz 4 Vr NGC 985 1=- £3- .2 J L J L J I I \ I I I I I I I I I I I I -.2 .2 .4 .6 .8 H - K (mag) •'igiin' 2.10 (('oiitiiiiicd) 65 rings, except for LT -H and knot 5 in V'll Zw 466. Knot -5 in V'll "Lw 466 is in a region of the ring where there is strong 15 fxm dust emission (.A.ppleton et al. 1997). LT 41 may also be suffering from dust emission, which would explain the red colors of the knots. ISO ob.servations of LT 41 are needed in order to determine the significance of dust emission in the ring. Finally, further indication that we have properly corrected the colors of the knots in II Zw 28 and .-Vbell 76. which suffer Balmer absorption is the fact that the ages of the knots, deduced from the H - K versus .1 - H color-color diagram (~6-8 Myr). are in excellent agreement with the other, mainly optical, color-color diagrams (~4-8 Myr). The knots in LT 41 are generally the reddest of the sample, and one knot is consistently redder than the models predict. What else besides dust emission or underestimates of .Vv may be the cause? IMF variations may play a role in reddening the colors of an evolving star cluster. Figure "2.11 is the same as Figures 2.8-2.10 but shows the effect of steepening the slope of the IMF. .-V comparison of Figure 2.11 with Figure 2.S revealed that the colors predicted by the LH9o n\odels with a Salpeter L\IF are redder than those predicted by the LH95 models with a steeper IMF (o = 3..50). .-V star cluster with a shallow IMF (i.e.. o < 2..J5) will have more massive stars than a cluster with a steeper IMF (i.e. a > 2..'}5). The excess of massive stars will pump the luminosity of the nebular lines higher than if the cluster had fewer high mass stars, which in turn leads to redder broad-band colors. Thus, if the I.MF of the star forming knots in LT 41 were, for some rea-son. shallow or "top heavy" (i.e.. had a large low-mass cutoff of > 1 .\I . ). the colors might agree with the LH9o models. However. .\I.\ revealed that LT 41 doesn't have an unusually large Ha luminosity, as it should if the knots in LT 41 contained more high mass stars, relative to the knots in the other rings. Therefore it is unlikely that the IMF in LT 41 is significantly different from the rest of the rings in the sample. If LT 41 is a weakly bursting ring it will appear redder than a coeval strong burst, since the red light from an underlying, older stellar population would be dominate. We show in Figure 2.12 the effects of varying the burst strength on the B - V and \' - K colors in the BC9") models. Overlayed are the B - \" and V - K colors of the knots in the rings. In making this model wc have assumed that the age of the underlying population is 15 Gyr. The knots in LT I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I Ih. Z 0.25 Zo Ih. Z Z© a = 3.30 ^ 1 Myr • 4 Myr • 10 Myr • 100 Myr .5 O) n) E, > I CQ / n o 05 .A* -.5 I I I I I I I I I I I I I I I < I I I I I I I I I I I I I I I I I I I I I I I -1 -.5 1 1.5 2.5 3.5 V - K (mag) l-'iffiirc 2.11 IJI!).') model,s, with a .slccixMicd IMI-", such that Ay = 1 mag J_L -1 -.5 0 .5 1 1.5 2 2.5 3 3.5 V - K (mag) l-'igiirc 2.1 1 (('oiilimu'ii) A 1 Myr b = 0.1 • 20 Myr b = 0.01 • 80 Myr b = 0.001 • 600 Myr CDto E > I CQ -.5 -1 5 0 .5 1 1.5 2 2.5 3 3.5 V - K (mag) l-'igiirc 2.12 'I'liis ligiirc coiilaiii.s tlic biir.si striMiglli mollified BCOrj models plotted on a \' - K versii.s H - V color-color diagiani (again witlioiil and with data). The plot with data is an overlay ol the models with the reddening correc ted \' - K versn.s H - \' colors of the knots in the rings, 'I'lie age of (he underlying stellar population is 15 (iyr. We show several lime ste|)s in the evolution of the models (1, 10, 20, 10 (iO, KO, 100, 200, 100, (iOO, SOO, 1000, TjUOO, lOOOO, and l.'iOOO Myr) by using geometrical symbols; live pointi>tl stars for tlie I) = 1 (.solid line) model, pins signs lor the I) = 0.1 (dot-dashed line) model, crosses for the I) = 0.01 (clashed line) model, and tliree-|)ointe(l stars lor the b = 0.001 (dotted line) model. • WN 1 A IIZw28 Ay = 1 mag • VII Zw 466 • Abell76 O Arp 10 • LT41 • II Hz 4 •k NGC985 -K .bH -*-x' I I I I I I I I -1 -.5 0 .5 1 1.5 2 2.5 3 3.5 V - K (mag) l-'igurc 2.12 (("uiitiiiiu'd) I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 1 - b = 1 A 1 Myr b = 0.1 • 20Myr b = 0.01 • 80 Myr b = 0.001 • 600 Myr O) CO E 00 -1 o -.5 I I I I I I I I I I I I I I I I I I I I I I I -1 -.5 .5 1 1.5 2.5 3.5 V - K (mag) l-'iffiirc "i.l.J lie .saiiic ii.s imguic 2.12, cxccijI lliat llic age of the niiilcrlyiiig .sicllar popiilalioii i.s 1 (lyr iu.sU'acl oC 15 (Jyi. • WN 1 A IIZw28 Ay = 1 mag • VII Zw 466 • Abell76 O Arp 10 • LT41 • II Hz 4 ^ NGC985 HH J_L I I I I I I 1_1 -1 -.5 0 .5 1 1.5 2 2.5 3 3.5 V - K (mag) l-'if^iirc 2.i;{ (("oiiliiiiu'il) 72 41 are best fit by a young (~10 to 60 Myr). but weak burst (~1 to lO'/c). The rest of the knots in the other ring galaxies are consistent with young, powerful bursts (~10 to lOO'X). Note that if the age of the underlying stellar population is 1 Gyr instead of 15 Gyr. as shown in Figure 2.i;{. the conclusions about the age and burst strength of the knots is not significantly changed, except that one knot in LT 41 is no longer fit by the model. .Although we cannot rule out dust emission or underestimates of .\v. LT 41 may be a weakly bursting ring with an underlying stellar population between I and 15 Gyr. 2.7 Summary and Conclusions Based on the spectra and optical/infrared broad band photometry of individual starburst knots in the ring galaxies, we conclude: 1) The nitrogen abundances show a correspondence with the diameter of the rings, with .\'/H spanning a range from appro.ximately one-fifth solar for the smaller rings up to one-half solar for the larger rings. These abundances probably reflect the original abundances of the target galaxies before the collisions. 2) The rings have approximately one-half solar o.xygen abundances, independent of the size of the rings. •i) If we assume there is a morphological segregation of Hubble-type with ring size, the smaller rings inhabit small late-type galaxies and the large rings inhabit larger mid- to early- type gala.xies. then the oxygen and nitrogen abundances are in the "normal" range of Hubble- type galaxies. 4) .\ simple theoretical calculation suggests that there may be in aicu production of o.Kvgeri in the ring waves. The calculation revealed that more o.xygen might be produced than we observe, in which case the o.xygen has been either be blown out of the disk, or produced and hidden in a hot. unobservable 10'' K phase. 5) II Zw 28 exhibits significant Balmer absorption in its spcctra. indicating the existence of a substantial post-starburst. .\-star population. It is possible that the .Vstars exist in a bar like structure seen in recent ILST observations. The velocity of the absorption lines is seen to 73 vary systematically across the bar. Abell 76 suffers from Balmer absorption in a single knot, which was speculated to have been part of the companion galaxy seen in projection against the ring. 6) The kinematics of LT 41 are consistent with a simple model where the ring component is simultaneously expanding and rotating. 7) Comparison of the colors of individual starburst knots with numerical stellar evolution models revealed that the knots have ages tliat lie in the range from 4 to SO .\Iyr. These ages are less than the dynamical ages of the rings (typically 100 to 300 Myr). 8) The fact that the knots from individual ring galaxies cluster around the same, very young age argues that their formation is recently triggered by a coherent event all around the ring. This strongly supports their production as being due to the propagation of a radial density wave. 9) LT -11 may be a weakly bursting ring with an older underlying stellar population between 1 and 15 Gyr in age. We wish to thank C. Struck for helpful comments, suggestions, and conversations. Bransford would like to thank Dr. Evan Skillman for his valuable suggestions and a stimulating conversion while visiting the L'niversity of Minnesota. We also thank D. Garnett for some helpful suggestions in the early stages of the project. This work is supported by .\SF Grant .\'o. AST-9319596. A. .\larston was supported by N ASA .JOVE grant .\.-\.G8-264 and a grant from X.-VS.A. administered by the .American .-Vstronomical Society. 2.8 References .-\.lon.so-Herrero. .-V.. .\ragon-Salamanca. Zamorano, .1.. 1*0 Rego. M. 1996. M.VR.-VS. 278. 417 .-Vppleton, P..\'.. i: .\larston. .A..P. 1997. .\.I. 113. 201 (.\M. Paper I) .Appleton. P..\.. Struck-.\larcell, C. 1996. in Fundamentals of Cosmic Physics. 16. Ill .\ppleton. P..\'.. Charmandaris. V.. i; Struck. C. 1996. .-Vp.!. 468. 532 74 Appleton et al.. 1998. in preparation Barnes, -J.V'.. Hayes. D.S. 1984. IRS Standard Star Manual. Revised February 1984. Kitt Peak .\'at. Obs. Publ. Bruzual. G.A.. &; Chariot. S. 1995. preprint (BC95) Charmandaris. V.. .\ppleton. P.N. 1996. .A.p.J. 460. 686 Cliarmandaris. V'.. .Appleton. P..\'.. L- Marston. .A.P. 1993. .A.p.1. 414. 154 De Veny. .1.. Carder. E.. ic Harnier. D. 1995 The Revised Gold Camera L'ser's Manual. 1995. Kitt Peak .\at. Obs. Publ. Dopita. L Evans. I..\. 1986. .A.p.J. 307. 431 Edmunds. M.G.. Pagel. B.E..J. 1984. .\I-\R.\S. 211. 507 Gao. Y.. Solomon. P.M.. Downes. D., Radford. S..J.E. .\p.IL. 481. L35 Garnett. D. 1990. .-Vp.!. 363. 142 Higdon. .I.L.. Rand. R..J.. Lord. S.D. 1997. .\p.JL. 489. L133 Horellou. C.. Casoli. F.. Combes. F.. iSj Dupraz. C. 1995. .-Vic.-V. 298. 743 Kobulnicky. Skillnian E.D. 1997. .-Vp.J. 489. 636 Leitherer. C.. Heckman. T..\I. 1995. .\.p.IS. 96. 9 (LH95) Lynds. R.. Toomre. .-V. 1976. Ap.J. 209. 382 Maeder. A. 1990. Ac^.-VS. 84. 139 Maeder. .-V.. fc .\Iaynet G. 1988. .-V&;.\S. 76. 411 Marcum. P.M.. .Appleton. P..\.. L Higdon. .1. 1992. .\p.I. 399. 57 .Marston. .A.P. Bransford. M..\. 1998. in preparation .Marston. .-V.P.. .Appleton. P..\. 1995. .AJ. 109. 1002 (M.A. Paper 11) Martin. P.. Belley. .1. 1996. .Ap.J. 468. 598 McCall. .M.L.. Rybski. P..M.. Shields. G..A. 1985. .Ap.JS. 57. I Osterbrofk. D.E. 1989. in .Astrophysics of Gaseous .\ebulae and .Active Galactic .Nuclei, pub. L'niversity Science Books 1 o Pagel. Edmunds. .\I.G.. Blackwell. D.E.. Chun. M.S.. k Smith. G. 1979. M.N'RAS. 189. 95 Rodriguez Espinosa. J.M.. ic Stanga. R..\[. 1990. Ap.J. 36-5. 502 Sargent. VV.L.VV. 1970. .A.p.J. 160. 405 Savage. B.D.. &: Mathis. .J.S. 1979. .•VR.A.ij.A. 17. 7:3 Skillman. E.D.. Kennicutt. R.C. 1993. .-Vp.!. 411. 655 Theys. .J.C.. i; Spiegel. E..A.. 1976. .-Vp-J. 208. 650 Thompson. L..A... Theys. J.C. 1978. .\p.J. 224. 796 Toomre. -V. 1979. in The Large Scale Structure of the Universe. I.\L" Symposium .\'o. 79. edited by .M.S. Longair and .1. Einasto (Reidel. Boston), p. 109 V'Ua-Costas. M.B.. Edmunds. \LG. 1993, MNR.-\S. 265. 199 V'ila-Costa.s. .M.B.. (k: Edmunds. .M.G. 1992. M.\R.A.S. 259. 121 Woosley. S.E.. Weaver. T..-V. 1995. .-\.p.IS. 101. 181 Zaritsky. D.. Kennicutt. R.. Huchra. .J.P. 1994. .A.p.J. 420. 87 (ZKH) 76 CHAPTER 3 NEUTRAL HYDROGEN OBSERVATIONS OF ARP 118 3.1 The Arp 118 System The Arp 118 system is comprised of a distorted disk galaxy NGC 1144 and an elliptical galaxy NGC 1143. NGC 1143 is ~40" to the .\VV of .\GC 1144. which corresponds to a linear separation of ~ 20 kpc (assuming a distance of 111 Mpc to .\rp 118. based on Ho = 75 km s~' Mpc"'). .\GC 1144 is also classified as a Seyfert 2 galaxy (Hippelein 1989. hereafter H89). There exists a knotty "ring" or loop, that extends from NGC 1144 towards the companion in the north-west. .Just inside the north-eastern edge of this loop, and somewhat above the brighter nuclear region, is a .second loop or filament. Figure .'J.l contains an Ha (-t-continuum) image taken with the KP.NO 2.1-meter telescope. In Figure .•{.2 I show an HST image (taken with the Ff)06V\' filter, image courtesy of .Matt Malkan) of the brighter nuclear region in the south-eastern half of .XGC 1144. .Just west of the nucleus is a bright extra-nuclear knot, shown by .Joy Ghigo (1988. hereafter .IG88) to be a site of vigorous star formation. .A. sharp dust lane can be seen running from the SE to the XW just east of the nucleus. The southern end of the dust lane appears to curl, bonding to the SW. .-Vlong the southern edge of the disk appears to be another "loop" (or filament), which runs appro.ximately east-west, curling up at its ends and joining the with the loop that stretches .\\V towards the companion. .IG88 revealed that the extra-nuclear star forming regions in the SE region of .\GC 1144 are very luminous at 10 //m. with the bright knot to the west of the nucleus contributing nearly H-VX of the total emission. .IG8S calculated XGC 1144 has a bolometric luminosity of 2.5 X 10" L,. . 80% of which is re-radiated in the thermal infrared. Their 6 and 20 cm radio continuum images revealed several radio synchrotron sources corresponding with blue optical emission knots, suggesting that the radio continuum was mainly coming from recent Figure 3.1 Hor image, taken with KP.VO "i.l-m telescope Figure :i.'2 Hubble Space Telescope image. Courtesy of .Matt .Malkan. 78 star formation. Condon et al. (1990) noted a slightly elongated radio source at the nucleus, with e.xtended emission to the .\E. Ghigo (19S3) has found the nuclear radio .source to be double with a separation of ~2". The high far infrared-to-blue luminosity Lf[r/Lb (~-i) and high color temperature Sioo^m/SeoMm imply strong star formation (.\ppleton Struck-.\Iarcell 1987). .\rp 118 has been found to have very high internal velocities, both in Ha (H89) and CO emission (Gao. Solomon. Downes. Radford 1997. hereafter GSDR). Hippelein obtained spec tra of H H regions distributed around the ring and found that their radial velocities vary by ~1100kms~'. Further. H89 calculated dynamical models of the encounter between .N'GC ll-t-l and .\'GC 114.3 which indicate that the disk of .\'GC 1144 may be rapidly rotating (~ 700 km s~'. or a rotation period of 88 million years in the "20 kpc diameter ring). GSDR detected CO emission which also covered a velocity range of ~1100 km s~'. GSDR found .\rp 118 to be an e.Ktremely luminous CO source, with a CO luminosity nearly twice that of .\rp 220. Their IR.\.\I single dish observations (22" beam) showed that the CO emission is distributed non-uniformly. and is highly concentrated in the SE quadrant of the ring. The single dish observations also showed CO emission emanating from the .WV quadrant of the ring. CO maps made with the [R.\M interferometer x 2'.'5 beam) revealed that the CO emission is not centrally concentrated near the Seyfert 2 nucleus, as is common in interacting gala.xies. but trace the .southern arm of the radial ring wave. In particular, the strongest knots of CO emission are coincident with the most luminous H II regions (which are also concentrated along the southern arm of the ring). 3.2 Previous HI and CO observations of Ring Galaxies HI observations of the Cartwheel and .\rp 143 (Higdon. Rand, i.: Lord 1997; Higdon 1996: .\ppleton. Schombert. tVj Robson 1994) revealed that ~90% of the HI emission is concentrated in the ring waves (see .\ppleton ic Struck-Marcell 1996 for a comprehensive review of collisional ring galaxies). HI observations of collisional ring gala.xies have also revealed that some systems may be recent mergers (e.g. .-Vrp 10. Charmandaris .\ppleton 1996) and that others have HI 79 plumes connecting the target and intruder galaxies (e.g. V'H Z\v 466. .-Vppleton. Charmandaris. k Struck 1996: Arp 143. .\ppleton. Schombert. .k Robson 1994). Until recently, the molecular content of this particular class of interacting gala.xy was not known. Horellou et al. (1995) found that ring gala.xies are bright in CO emission relative to non-interacting galaxies, and the CO emission is mainly centrally concentrated, consistent with CO observations of other interacting galaxies. Higdon. Rand, Lord (1997) have made high resolution CO images of collisional ring-like system .A.rp 143. These observations revealed the CO emission is centrally concentrated (> 80% of the CO emission is from the nucleus), however two strong CO emission knots, from putative giant molecular cloud associations, were found to e.xist in the extra-nuclear ring wave (Higdon et al. 1997). Spatially coincident with these GM.-Vs are several luminous H II regions and a banana-shaped ridge of HI emission (.see .-Vppleton. Schombert. Robson 1994). In this chapter we present our (P..\'. .-Vppleton and K. Freeman) VL.-V HI observations of .-Vrp 118. in conjunction with the CO observations of GSDR (the CO data was generously provided by Yu Gao). We organize this chapter as follows. In section .3.3 we present the details of the HI ob servations and data reduction. In section 3.4 we present the global HI profile and the HI distribution of .-Vrp 118. .-Vdditionally, we di.scuss the CO emi.ssion. and its relation to the HI emission. Section 3..t contains the presentation of the HI kinematics. In section 3.6 we discuss possible interpretations of the HI observations. In section 3.7 we present the HI distribution and kinematics of an Hl-rich dwarf companion to .-Vrp 118. Finally, in .section 3.8 we present our summary and conclusions. We assume throughout this chapter a distance to .-Vrp 118 of 111 .\Ipc, based on a heliocentric velocity of 8288 km s~' (based on these HI observations) and a Hubble constant of 7-5 km s~'. 3.3 VLA C-Array HI Observations of Arp 118 .-Vrp 118 was observed at the \'L.-V on 17 .March 1996 using all 27 telescopes configured in the C-array. The correlator was set in a two IF (intermediate frequency) mode (2.'VD) with on-line 80 Manning smoothing and :j2 channels per IF. For each IF we used a bandwidth of 6.25 MHz. which provided a frequency separation of 195.3 kHz per channel. This frequency separation corresponds to a velocity separation of 43.8 km s~' channel"' in the rest frame of the galaxy (using the optical definition of redshift). .\rp 118 contains HH regions (in the ring) that have radial velocities spanning a range of more than 1000 km s~' (H89). GSDR reported a wide velocity range in (1100 km s~') CO emission. In order to achieve a velocity coverage consistent with the observed velocity range. IFl was centered at 8269 km s~' and IF2 at 9230 km s~'. The resulting velocity coverage of our observations was 2275 km s~'. .A. total of 3 hours and 47 minutes were spent on source. Strong winds during the observations were responsible for a loss of 25% of our observing time. FIU.K and phase calibration w-ere performed using the sources 3C 48 and 0320+053 (B1950). respectively. The data were first amplitude and phase calibrated, and bad data due to interference were flagged and ignored by the .-VIPS software. .\n image cube was created from the L'\' data by giving giving more weight to those baselines that sampled the L'\' plane more frequently (natural weighting). This provided a synthesized beam with a FWHM of21''2 x 17'.'0 for both IF 1 and IF2. Subtraction of the continuum emission in each line map was performed using a standard interpolation procedure based on five continuum maps free from line emission at the ends of the bands. The rms noise per channel was 0.36 m.Jy beam"'. The highest dynamic range achieved in any channel map was 10:1. .-V single image cube (we will refer to this as the combined cube) was made by combining IFl and IF2. In order to determine the total HI distribution, we employed the following technique. We smoothed the combined cube with a beam twice that of the synthesized beam. .\s a result, a smooth cube was created WMth a resolutioti of 42'.'4 x ."M'.'O. We applied a signal-to-noise threshold to the smoothed cube, blanking any pi.xels that fell below a 3 produced by adding tlie new maps together, and the same technique was used to create the first and second and moment maps of the distribution. Using this procedure we effectively give more weight to points associated with low surface brightness emission. 3.4 Integrated Properties, HI Distribution, and the Relationship Between the HI and CO Emission Figure 3..'} contains the global H[ profile of .A.rp 118. Notice the narrow spread in velocity of the emission, spanning ~ km s~'. There are faint absorption features at high velocities, which we discuss later. The global HI profile has about one-third the velocity spread of the CO and Hfv emission, as shown in Figure .'{.-L .\ majority of the CO emission is at high velocities (in the SE) whereas and the Hi emission is at low velocities (in the .WV). The HI and CO velocities are in appro.ximate agreement with the Hcv velocity field, which revealed that the H II regions were at low velocities in the .WV quadrant of the ring and at high velocities in the SE quadrant (H89). We derive a heliocentric velocity V'hi of 8288 km s~'. which is .{fiO km s~' below the optical velocity of .\GC 11-14. Conveniently, the total HI mass \Ih may be calculated from the 2l-cm emission spectrum (i.e. the total HI mass of .-Vrp 118 may be calculated from Figure This procedure is outlined as follows. Consider a gala.vy with a surface density of hydrogen gas. n atoms cm~". emitting the 21-cm line. Further, iissuine the gtis is optically thin such that the photons can easily escape. Therefore, there is a direct relation between the number of photons emitted and the number of hydrogen atoms along a particular line-of-site. Simply put. the probability of a hydrogen atom emitting the 21-cm line multiplied by the total number of hydrogen atoms in a gala.Ky is equal to the total flu.\ received (the total flu.x being the area under the curve in Figure for example). The surface density of hydrogen radiating in a 1 km s~' interval can be written (e.g. Wright 1974); n = 1.823 X 10^^ Tb (atoms cm ^) (3.1) where Tb is the brightness temperature (units of K) of the line emitting gas. and i.s related to 15 10 Typical errorbar '-eJ 5 0 — •5 i— 8000 8200 8400 8600 8800 9000 9200 velocity (km s ') I'igiiro (ili)l)}il III profile of Ar|> IIN. Tiic arrow.s iK)iiil to absorption (caliirc.s. riic faint al>Morplioii fcatiirc al km s"' will need riirllicr olwiMvalion.s to confirm, licncc the (pic-slion mark. 1.5 ~i—I—r "1—I—f- -|—I—r "I 1 r "1 r "I—I—r -|—I—r range of velocities from Ha in H II regions Single dish CO emission of NGC 1144 w I r:; -1,1 c I I D « I loI < 'I '• -i" J I L J \ L J L J I L J i. J I L J ! L 8000 8200 8400 8600 8800 9000 9200 9400 velocity (km s~^) •'igiirc I (Il()l)al ill profile of Arj) IIH (.solid line) over-plotted l)y glolwl profile of ('O emission (dashed line, (i.SDK). The the fliix density of the source (as shown below). A standcard way of calculating (e.g. \erschuur 1974: Wright 1974) the total number of hydrogen atoms .\. assuming small optical depth, is to integrate the surface density, n. across the gala.xy and over the velocity distribution of the line-emitting gas: N = X f d.\ f TB (atoms) (:}.2) ^galaxy V velocity where d.\ is a differential element of the surface area of the gala.xy. and is related to the distance D by d.-V = dfi (dQ is the differential solid angle subtended by d.\). and d\' is a differential element of the velocity distribution in km s~'. Substituting for d.\. we get: .\* = l.82:{ X lO*'^ J J Tb dV dQ. (:}.;}) .-Vt 21-cm hi/ kT. hence the Planck function can be appro.Kimated by the Rayleigh-.Ieans radiation law. such that Tb is related to the flu.K density through the following relation: Tb =|^ (K) ,:u) where u is the frequency of the 21-cm line (in kHz), Si, is the flux density (in .ly). c is the speed of light, and k is the Boltzmann constant. The HI mass is the product of the total number of atoms and the mass of a single hydrogen atom. Substituting equation (."{.4) into and multiplying by the mass of a hydrogen atom, the total HI mass is (in solar units): ^ = 2.356 X 10" D- J S:, dV (;}..5) where / Sj, d\" is often referred to as the integrated flux density. The units of the integrated flux density are .Jy km s~' since S^ is in .Jy and dV is in km s~'. The integrated flux density detected by Bushouse (1987) using the .Vational Radio .-Vs- tronomy Observatory (.\R.-\0) 91 m telescope was 2.8.'} Jy km s~' and by .leske (198(J) using .\rccibo was 2.oo .ly km s~'. Our interferometer observations detected 2.40 .ly km s~^ or 85'X of the total cmis.sion quoted by Bushouse and 94*/( of the total emission quoted by .leske. Our 8.5 profile is in general agreement with those of both Bushouse and Jeske and is missing and respectively, of the flux observed in the single-dish measurements. This probably results from the fact that the C-configuration of the VLA has resolved out some of the more extended emission. The total HI mass of Arp 118 based on our C-array observations is Mh = 6.97 x 10^ M ;. In Table ^J.l we present the derived HI properties of .\rp 118 (and its companion 7f8 to the north-east: the companion will be discussed in section 3.T). We also include in Table ."i.l an estimate of the dynamical mass (see .-Vppleton et al. 1996) of .-Vrp 118 and its companion. The values of M,i were calculated using the formula: [(cosec(i)]'^ RHI Md = —— (M,) (;}.6) where Rhi is the radius of the HI disk. AV'y, is the width of the profile where the [lu.x ilensity is one-half the peak, and 1 is the inclination of the disk. This formula applies to gas that is in bound circular orbits. .\ote that the assumption of circular orbits and a "normal" rotation curve may be quite incorrect in .A.rp 118. Considering the fact that the spread in velocities of the HI emission is about one-third that of the CO (and Ho), we have estimated -^V from the combined HI and CO profiles, giving a of 800 km s~'. Because we employed the combined CO and HI profiles to estimate the dynamical mass, we used the radius of the ring. 10 kpc (calculated assuming a distance of III Mpc to .\rp 118 employing a Hubble constant of 75 km s~' Mpc"'). in equation (3.6). We employed a value of i = 50° for .-Vrp 118. We assume i = -15° for the companion, since the optical "disk" is unresolved (the companion "gala.Ky" is actually an HST guide star! See section 3.7). Figure 3.5 contains the integrated HI distribution of .A.rp 118. overlayed with a digitized sky survey (DSS) grey-scale grey scale image of .-Vrp 118. The HI emission is distributed non- uniformly throughout the disk of N'GC 11-14. is somewhat concentrated in the XW quadrant of the ring, and contains a protrusion of emission to the south. Figure 3.6 (see also Figure 3.7. below) is an overlay of the integrated HI map with the a grev-scale image of the CO knots imaged bv GSDR. .-\s mentioned, the high resolution CO 86 / V 2400 02 5242 40 38 36 34 RIGHT ASCENSION (B19SQ) Figure 3.5 Grey scale DSS image of Arp I IS with a contour map of the integrated HI distribution. The contour increment is 50.0 .ly beam"' ni and the level of the lowest contour is 100 .Jy beam"^ m s~'. 87 -Q02215 30 lO 45 at \ » t r- m f ' ^ ' i z oO 23 00- z locadon o Ui 15- • 30- most intense CO emissian. 45- CO knots einn^y a smaller beam size. 2400- 02 52 42 41 40 39 38 37 36 35 34 RIGHT ASCENSION (B19S0) Figure Integrated HI contour map overlayed with a grey-scale image of the CO knots imaged by GSDR. The CO knots are at a different resolution that the \'LA beam. The [RAM interferometric CO observations conducted by GSDR cm- ployed a o'.'.i X 'Z'O beam. 88 -002215- 30- s- us a» 45- m z 9 23 OOh s CO emissiaa canvotved o to VLA be2aii UJ 15- • 30- 45- 2400- 1 1 1 1 02 52 42 41 40 39 38 37 36 35 34 RIGHT ASCENSION (B1950) Figure 3.7 Same as Figure .3.6. e.xcept the CO emission is convolved to the same \'L.\ beam shape used in the HI observations. 89 Table ;}.l HI properties of Arp 118 and its companion Property" .Arp 118 Companion a(Bl9oO) 02^ 52'" 37!6 02*^ o;r 1.3M d'( 819.50) -00° 22' ol'.'S -00° 18' 26''2 V'hi (km s-') 8288±45 8749±45 AVi/^ (km s"') 204 51 AVy. (km s~') 375 141 Rhi (arcsec) 16 10 /S(V')dV' (.Jy km s~^) 2.40 1.06 .MHI/M.. 6.97x10^ 3.42X 10'' Md (M. ) [6.34x10^1]^ 1.61x10^ "Linear distances at the sources are calailated assuming a distance of 111 Mpc to Arp 118. and 117 Mpc to the companion, employing a Hubble constant of 75 km s~' Mpc"'. ""Computed using c-md the radius of the disk. See te.tt. / 2 data of GSDR shows that the CO emission traces the southern arm of the radial ring wave (the appro.ximate location of the ring has been drawn to guide the eye). Further, the knots of CO emission are coincident with luminous H II regions, also concentrated along the southern arm of the ring. Inspection of Figure 3.6 reveals that the brightest comple.xes of CO emission do not overlap the HI emission. In Figure 3.7 we have convolved the CO emis.sion to the same beam size as our \'L.\ data. The CO emission appears to. more strikingly, "fill-iir the "missing" HI eiiii&sion in the south-eastern half of .\'GC 1144. Considering Figures 3.4. .{.6. and 3.7. the HI and CO appear "segregated", both spatially and kinematically. Xote. however, that we use the word "segregated" rather loosely, since there is weaker CO emission in the .\VV quadrant (see ne.Kt paragraph), and low-level HI emission in the SE. The word is used to point to the fact that the distributions of CO and HI are asymmetrically peaked. CO to the SE and HI to the .\\V. Interestingly, there is a striking similarity between the single dish CO profile in the .\\V quadrant of the of the ring and our HI profile (see Figure I in GSDR for the single dish profile). The CO profile peaks at a velocity of S'2.50 km s~' and the HI profile peaks at 8267 km s~'. Both profiles fall off rapidly towards higher velocities. W'e have noted that the CO and HI have asymmetric distributions. .-V more quantitative 90 8617 8S73 -002200^ 30 r 2300H i 2400h \-^r 8S30 8442 -0022 00^ '] i 'Vi 3or c; i 2300j^ 30h 2400h S389 83S5 8311 2-0022 00: Z o 20\ t 2300- -1 u 30h Ul a r 2400- 8267 8224 8180 -00 22 00 r 30h 2300h 30^ /O ^•s 2400r r 02 5242 40 38 36 34 02 52 42 40 38 36 34 RIGKT ASCENSION (B1950) lire ;}.8 Cotitour plots of the 12 channel maps, overlayed with the DSS gray-scale image of Arp 118. The velocity (in km s~') is displayed in the upper right. The contour increment is 3.6 x 10"' .Jy beam"', and the lowest contour displayed is at a. 91 description of the relative importance of the atomic and molecular hydrogen content across the disk N'GC 1144 would be to calculate the molecular-to-atomic gas mass ratio. -MHi/Mni- i" the .\VV and SE regions. We have converted the observed CO emission from GSDR into a mass of molecular gas (using the standard galactic conversion from CO to H2). and calculated the HI mass, separately for the .\W and SE regions. The molecular-to-atomic gas mass ratio is 0.77 in the .WV section of the ring, consistent with the observed ratios in normal spiral galaxies (Casoli et al. 1998). The ratio is very large in the SE section. 17.6. consistent with ratios measured in other interacting, infrared luminous galaxies (Mirabel Sc Sanders 1989). The gas mass ratios suggest that the dominant state of the interstellar medium proceeds from mainly atomic in the S\V to molecular in the SE. In the SE region less than 6% of the interstellar gas is in atomic form. Considering that the total (HI-fH)} mass of hydrogen remains approximately equal in both regions (~L and 2 x 10'° Min the N'W and SE. respectively) strongly suggests that there is a large scale conversion of HI to H2 in the southern region of N'GC 1144. We discuss this possibility further in section 3.6. 3.5 HI Kinematics In Figure ;{.8 we present the HI channel maps. In the rest frame of the gala.vy. the channel separation is 43.8 km s~'. The emission starts to increase in strength around 8;j.>5 km s~'. ancl is subsequently seen to be coming from a series of knots which becomc stronger, peaking near a velocity of 8*267 km s""'. The centroids of the emission appear to shift from just north-east of the ring at a velocity of 8.J11 km s~^ through the north-western tip of the ring (whore the emission is strongest). .\t a velocity of 8180 km s~' the emission shows two streamers of HI. both in the ring and just to the south-west of the ring, pointing back towards .\"GC 1143. These features could be accretion streamers and/or gaseous bridges as observed in \'Il Zw 466 (.-Vppleton, Charmandaris. <*c Struck 1996) and modeled by Struck (1997). Figure 3.9 contains the mean velocity field of Arp 118. which appears rather disturbed. The velocities appear, appro.xiniately. to run from lower velocities in the north-west to higher velocities in the south-west. E.xcept for contours at 8280 and 8300 km s~^ there is little 92 emission from the SE. The velocity isocoritours generally have a bent and/or rather distorted S-shaped appearance, suggesting the disk of NGC 114-1 is very warped. Figure .J.IO contains another set of channel maps showing the faint absorption features, probably against the Seyfert 2 nucleus of .\GC 1144. The absorption is strongest at a velocity of 9011 km s~'. but is very faint in the other channels shown in Figure 3.7. Considering only the strongest absorption feature, we derive an HI column density (see. for e.vample. .Mirabel (1982)). in terms of the spin temperature. Ts. of ~1 x lO'® atoms cm~^ Future observations of high sensitivity and with higher velocity resolution are needed to confirm the e.xistence of the absorption features. 3.6 Discussion One possible way to enhance the molecular gas mass fraction is by compressing the disk in XGC 1144. .Modest compression and/or shocks in the interstellar medium can lead to the conversion of atomic to molecular gas (Elmegreen 1993: Honma. Sofue. Arimoto 1995). The salient features of Elmegreen's model are that HI-H j gas phase transition depends sensitively on the pressure in the interstellar medium and the radiation field: H > molecules are formed on the surface of dust, but can be destroyed by L'V photons. The results of his models imply that large regions within gala.xies can spontaneously convert atomic into molecular gas following an interaction, a tidal encounter that has led to accretion, or increase in the gas surfacc density. Figure 3.2 (the HST image of .N'GC 1144) confirms that the southern .section of the disk of .\GC 1144 does indeed appear "squeezed". The nucleus of NGC 1144 is very off-center, which probably resulted from a highly off-center collision between .\GC 1144 and .\GC 1143. which has flung the nucleus of .\GC 1144 to the SE (e.g. see Figure 6 of Lynds and Toomre (1976)). The southern filament (shown in Fig. 3.2) appears flattened, and the long, sharp dust lane running .\W-SE also indicative a strong shock (possibly from the nucleus ramming into the IS.M at high speeds). C. .McCain (1997) has noted the e.xistence of a velocity discontinuity (or jump) between the nucleus and the surrounding gas. consistent with the possibility that the off-center motion of the nucleus may be driving large scale shocks throughout the SE region of 93 8.6 8.8 9.0 9.2 9.4 9.6 9.8 -00 22 30H- 45 DEC (B19 50) 23 00 8340^AM 15 \ 8380 8400 8420 8460 \ 8440 30 02 52 40 39 38 37 36 35 RIGHT ftSCENSIOH (B1950) Figure 3.9 Mean velocity field of Arp 118, with the velocity coiitours labeled in units of km s' ' 94 -00 22 00 9099 9055 9011 30 23 00- Q: 30- 0>, ; 2400^ lO oi t— -0022 00- 8967 8923 8879 m z tj o 30 z 23 00- Ij o 30- UJ • 2400 •y -0022 00- 8835 8791 30- 23 00- 30- 24 00- 0252 42 40 38 36 34 RIGHT ASCENSION (B19S0) F"igure .'{.lO Contour plots containing faint absorption features. The contour increment is -3.3 X 10""' .ly beam"' and the lowest contour displayed is 3 a. 95 NGC 1144. Tidal dissipation during the collision may also be responsible for driving gas into the nuclear regions, and we may catching .-Vrp 118 before niuch of the molecular gas has had time to sink into the nucleus. Radial motions of gas into the nucleus, coupled with the off- center motion of the nucleus, are probably helping to further squeeze the ambient interstellar medium. Interestingly, if the atomic hydrogen is somehow being converted to molecular form, the process of conversion must occur on timescales < one rotation period (88 million years), else the gas would have mi.ved (and/or sunk into the nuclear regions). The L'V radiation field from OB stars should increased the ambient radiation field, which should have the effect of converting molecular hydrogen back into atomic form (although the effects of star formation may be somewhat ameliorated by the fact that large amounts of dust usually imply higher metallicities. such that the L'V" radiation field may not be as intense, as in a less dusty environment). However. OB stars live for a few x 10' years, which doesn't change the fact that the conversion process, if operating in .N'GC 1144. must occur rapidly. 3.7 HI Distribution and Kinematics of the Dwarf Companion Figure .'i.ll contains the global HI profile of the companion (~8' to the XE of Arp 118). which is narrow and single peaked, with a of 51 km s~'. The HI mass of this system is 3.4 X lO''^ \l:. roughly half the value of Arp 118 (see Table 3.1). This "dwarr" gala.xy therefore appears to be somewhat rich in atomic hydrogen. The systemic velocity (V'nr) 'S 8749 km s~'. which gives a velocity difference of ~460 km s""' between the companion and .\rp 118. In Figure 3.12 we present the integrated HI image of the companion, overlayed with a DSS grey- scale image. Figure 3.12 reveals that the HI emission is co-spatial with a rather featureless, stellar looking "gala.xy". The HI emission is smooth, with a peak that appears to be somewhat off-set from the optical image of the gala.xy. The emission appears to be elongated to the SW (towards Arp 118). We have discovered that the HI emission from the companion corresponds to the same spatial location as the gala.xy KL'G 0253-003 (or PGC 011066). which was found by fakase iV; 10 — Typical errorbar 8 C) -e- •-8 6 — 0] C5co 2 — 0 — -e- J L_i- J 1 1 \ 1 1 L 8600 8650 8700 8750 8800 8850 8900 velocity (km s"') 11 (ilobal III |)i'i>lil(> of liwarf companion. -001730 19 00 30 02 5318 16 U 12 10 RIGHT ASCENSION (81950) giire ;M2 Grey scale DSS image of dwarf companion with a contour map of grated HI distribution. The contour increment is 25 Jy beam"' m the level of the lowest contour is 75 Jv beam"' m s~'. 98 • -001730 18 00 •» 30 19 00 30 • g-0017 30 •o % Ol ffl Z 18 00 O S z 30 -I o u • 19 00 30 02 5318 16 U 12 10 08 RIGHT ASCENSION (B195Q) Figure;}.!:} Channel maps containing HI omission From dwarf companion. The contour increment is 3.3 x I0~' .ly beam"' and the lowest contour displayed is 3 a. 99 8740 02 5318 16 14 12 10 08 RIGHT ASCENSION (B19S0) Figurp 3.14 Mean velocity field of dwarf companion, with the velocity contours labeled in units of km s~'. 100 Miyauchi (1988) in a survey of ultraviolet-excess galaxies. .\ED lists this galaxy as a "spiral" with a blue magnitude of 16.5. and with major minor a.xes of 0f3 x Of;3. The redshift of this gala.xy was unknown until this work (see above). There are no obvious spiral arms in the DSS image, hence the galaxy may be a dwarf spheroidal or elliptical galaxy, or it may be a very low-surface brightness disk galaxy with a relatively bright nucleus. Figure contains the channel maps with HI emission from the companion. .-Vlthough not very well resolved in velocity, the emission does seem to swivel about the "nucleus" of the galaxy, and hints that we may be seeing rotation. Figure 3.14 contains the mean velocity field of KL'G 02o.}-003. which also hints at rotation, with a major kinematic axis that runs approximately SE to the NW. Future observations with higher spatial and velocity resolution are needed to better understand the kinematics of this galaxy. Deep broad- and narrow-band imaging would be useful in order to determine if there is an optical counterpart to the HI disk. Spectra would prove useful in determining an optical redshift. and to determine the kinematics of the ionized gas (if any). 3.8 Conclusions The HI observations of .Vrp 118. and comparison of the HI and CO emission, reveal the following. 1) The HI emission, in addition to being highly disturbed, is distributed non-unifornily throughout the disk of .\GC I l-l-l. The HI emission is highly concentrated in the .WV quadrant of the ring. The CO emission is also distributed non-uniformly. and is highly concentrated in the SE quadrant of the ring. There is little or no CO emission in the nucleus, as is common in interacting galaxies. 2) The global HI profile has about one-third the velocity spread of the CO and Ha emission, majority of the CO emission is at high velocities (in the SE). whereas and the HI emission is at low velocities (in the .\'W). This is in approximate agreement with the Ha velocity field, which revealed that the H II regions were at low velocities in the .\'\V quadrant of the ring and at high velocities in the SE quadrant. Thus, the CO and PH emission appear to be "segregated". 101 both spatially and kinernatically. 3) High velocity HI gas is seen in absorption, probably against the (Seyfert 2) nucleus of NGC 1144. Although weak, the absorption appears to be spread over a velocity range of ~;}00 km s~'. 4) The moleciilar-to-atomic gas mass ratio. MH,/MHI- 'S 0.77 in the .\W section of the ring, consistent with the observed ratios in normal spiral galaxies. This ratio is very large in the SE quadrant. 17.6. and is consistent with the observed ratios in interacting, infrared luminous galaxies. These ratios suggest that the dominant state of the interstellar medium proceeds from mainly atomic in the NVV to molecular in the SE. In the SE quadrant, less than 6% of the interstellar gas is in atomic form. 5) The fact the total (HI+H2) hydrogen mass is approximately the same in the SE and the .\'W. but that the relative ratios of molecular-to-atomic mass change significantly, argue strongly that there is a large-scale conversion of HI-to-H ) taking place in he SE region of NGC 1144. This process probably occurs on a timescale < one rotation period (or 88 million years), which is rapid. 6) There exists a Hl-rich dwarf gala.Ky 8' to the .\E of .Arp 118. The emission is somewhat off-set from the optical image of the galaxy, and extends to the SW. The velocity field hints at rotation, but further observations will be needed in order to better understand the kinematics of this gala.KV. 3.9 References •Appleton. P.X.. Struck-Marcell. C. L987. .\p.I, 312. 066 .\ppleton. P..\.. Struck-Marcell, C. 1996. in Fundamentals of Cosmic Physics. 16. Ill Appleton. P..\.. Charmandaris. V.. Struck. C. 1996. .A.p.1. 468. .532 .-Vppleton. P..\.. Schombert. .J.M.. &: Robson. E.I. 1992. .\p.l. 385. 491 Bushouse. H..\. 1987, .-Vp.J. 320. 49 Casoli. F.. Sauty. S., Gerin. \I.. Boselli. A.. Foucjue. P.. Braine. .1.. Gavazzi. G.. Lequeux. .1.. Dickey. .J. 1998. Ac^cA. 331. 451 102 Cliarniaciclaris. V'.. Appleton. P..\. 1996. Ap.I. 460. 686 Condon. .1..!.. Helou. G.. Sanders. D.B.. Soifer 1990. ApJS. 73. .•}o9 Elmegreen. B.C. 199."}. Ap.J. -HI, 170 Jeske. X.A. 1986. PhD Thesis. L'niversity of California. Berkeley .Joy. \[.. Ghigo. F.D. 1988. .A.p.J. 332. 179 Gao. Y.. .Solomon. P.M.. Downes. D.. Radford. S..J.E. 1997. .A.pjL. 481. L3o Ghigo. F.D. 1983. BAAS. 15. 6o8 Higdon. .I.L. 1996. .\pJ. 467. 241 Higdon. .J.L.. Rand. R..I.. ^ Lord. S.D. 1997. .A.p.IL. 489. L133 Hippelein. H.H. 1989. 216. 11 Honma. .\I.. Soiife. Y.. Arimoto. X. 199-5. Air.-V. 304. 1 Keel. W.C. 1996a. A.J. 111. 696 Keel. W.C. 1996b. .-Vp.JS. 106. 27 Lynds. R.. Toomre. .-V. 1976. .\p.J. 209. 382 McCain. C. 1997. PhD Thesis. Australia National L'niversity. Canbera Mirabel. I.F. 1982. .-Vp.I. 260. 75 Mirabel. I.F.. Sanders. D.B. 1989. Ap.IL. 340. L53 Struck. C. 1997. Ap.JS. 113. 269 Takaiie, B.. ic Miyauchi-Isobe, .\'. 1988. .\nnals of Tokyo .-Vstronomical Observatory. 22. 41 \erschiuir. G.L. 1974. in Galactic and Extra-Galactic Radio Astronomy. G.L. Yerschuiir K.L Kellermann (eds.). (.\e\v York. NY: Springer-Verlag). p. 27 Wright. M.C.H. 1974. in Galactic and Extra-Galactic Radio Astronomy. G.L. Verschuur K.L Kellermann (eds.). (.New York. .NY: Springer-\erlag). p. 291 103 CHAPTER 4 RADIO-LUMINOUS SOUTHERN SEYFERT GALAXIES. 1. RADIO IMAGES AND SELECTED OPTICAL/NEAR-INFRARED SPECTROSCOPY A paper accepted to The Astrophysical Journal M. A. Bransford. P. Appleton. C. A. Heisler. R. P. Norris. A.P. Marstoa 4.1 Abstract This is the first of two papers in which a study is made of a sample of 12 southern radio- luminous Seyfert galaxies. Our aim is to investigate possible correlations between radio mor phology and nuclear/circumnuclear emission-line properties. In this paper we present radio images at A = 13. 6. and 3 cm taken with the .\ustralia Telescope Compact .-Vrray (ATC.-V). global far-infrared (FIR) properties for the whole sample, and optical and near-infrared (MR) spectroscopy of an interesting subset. We find a mixture of radio morphologies, including lin ear. diffuse and compact sources. When the FIR colors of the gala.xies are considered there is an indication that the compact radio sources have warmer FIR colors than the diffuse sources, whereas the linear sources span a wide range of FIR colors. There is a wide variation in radio spectral-indices, suggesting that free-free absorption is significant in some systems, particu larly IRAS 1 r2-l9-2859. .\GC 4507. and .\GC 7213. Detailed emission-line studies are presented of 4 galaxies IC 3639. NGC 5135. .\GC 3393 1*0 IR.-VS 11249-2859. In IC 3639 we present evidence of vigorous, compact star formation enclosed by very extended [01] A6300 emission, suggestive of the boundary between a diffuse outflow and the surrounding IS.\I. In another gala.xy. IC 5063. we see evidence for the possible interaction of a highly coUimated outflow and the surrounding rotating inner disk. Of the 5 104 galaxies which show compact radio emission. 4 have radio/FlR flux ratios consistent with an energetically dominant AGX. whereas IC 4995 exhibits evidence for a very compact starburst. 4.2 Introduction Indications that Seyfert gala,xies may contain a significant starburst component come from several different directions, although it is not yet clear how such star formation might affect the properties of an active galactic nucleus (.AG.N). .Vorris. .-Vllen. Sc Roche (1988) and Roy et al. (1997) showed that most Seyfert galaxies, unlike radio quasars, roughly follow the radio- far-infrared correlation, which suggests that their bolometric luminosity is dominated by star formation activity. Extended minor-a.xis radio emission from Seyferts (Baum et al. 1993) suggest.s that many Seyfert gala.xies may contain powerful circumnuclear starbursts. The inner few hundred parsecs of Mrk 477 have recently been shown to contain an e.xtremely luminous and compact starburst (Heckman et al. 1997). In N'GC 1068 an extended luminous starburst region is observed 3 kpc from the central .-VG.V (e.g. Telesco ^ Decher 1988). Genzel et al. (199-5) made high-resolution .\IR spectroscopic images of .\GC 7469. which show a circumnuclear ring of star formation 480 pc in radius. A blueshifted ridge of emission between the ring and the nucleus is interpreted as infalling gas that may provide fuel for the .\G.\. The star-formation in NGC 1068 and .N'GC 7469 may result from a stellar bar which is both triggering extended star-formation and funneling material inwards to the .\G.\. Previous investigations have shown that half of the total luminosity of .\GC 1068 (Telesco et al. 1984: Lester et al. 1987) and .Mrk 477 (Heckman et al. 1997). and nearly two-thirds of that of NGC 7469 (Genzel et al. 1995). arises from starburst activity. This study is aimed at looking for possible correlations between arcsec radio structure' and optical-lR emission-line properties in a sample of southern radio-luminous Seyfert galaxies (see Table 4.1). The mapping of arcsec radio structure in southern Seyfert galaxies has only become feasible recently with the advent of the .\ustralia Telescope .National Facility and many of these radio sources have not yet been mapped. Our pilot study focuses on one aspect of '.-\t the distance of our sample galaxies this corresponds to linear scales of a few hundred parsecs. lOo the properties of AGN's. namely the possible relationship between AG.\ or starburst activity and the morphology of the arcsec-scale radio emission. For example, we wish to investigate the possibility that radio-emitting plasma ejected from an .A.GN might interact and trigger star formation in the inner regions of galaxies, .\lternatively. in other systems, the radio emission itself may originate in circumnuclear star formation and this will be investigated via spectroscopic observation. In this paper we present the radio images, at three wavelengths, of the galaxies in our sample, as well as optical and .\'IR spectra for an interesting subset. In addition we will compare the thermal properties of the dust emission (as measured by global FIR properties of the galaxies, e.g. Heisler Vader 1994: Condon et al. 1991) with the radio properties. We show that for a subset of these data, the FIR color temperatures seem to correlate with radio morphology, suggesting that in some cases the FIR colors are intimately connected with the nuclear dust properties of Seyferts (see §-t.o.:3). In a second paper (Paper II) we will complete the spectroscopic observations and draw more general conclusions about the sample. In §4.3 we present the sample selection, in §4.4 the details of the observations, in §4.5 the radio data, and in §4.6 we present optical and N'lR spectra of some of the galaxies. In §4.7 we present our summary and conclusions. 4.3 The Sample The sample galaxies listed in Table 4.1 are a radio-bright subset of a 2.4 GHz .-Vustralia Telescope .National Facility (.A.T.\F) survey of Seyfert galaxies by Roy et al. (1997).^ Galaxies from this list were selected if they were more luminous than L^.-igHz > 10^'"' W Hz"', with declinations < -20°. and redshifts z < 0.04. One gala.Ky that met these criteria. IC 4:J29.V. was not included in our sample since it had been imaged in the radio by .-V.L. Roy (private communication) and radio maps of this source will be published elsewhere. Our 2.4 GHz observations reveal that the range in luminosity of our sample is log(L)..iGHz/^^-Hz~') = 21.7-23.2 (see Figure 4.1). By way of comparison. Centaurus .-V. the most radio-luminous "The Roy el al. sample is tlrawti from the cle Grijp et al. ID92 sample of 1R.\S selected galaxies. 'I'ablo 1.1 'Tlu' Sample Nil IMC HA (J 2000)" Di'X" (.12000)" mi.'' Hc'dsliilt'' Distaiici''' Scale'' 'lyije' Diaiiictcr'' (ll 111 s) (d III s) (mag) (M|,c) (k|)c/arcfiiT) (arciiiiii) N(;c m:i 10 IK 23.1 -IF) 09 11.0 13.09 0.0137 51.8 0.200 Sy 2 2.2 X 2.0 IH.AS ir2l5-2S0(i 11 21 02.7 -28 23 15.3 13.00 0.0135 51.0 0.202 Sy 2 .\ IH.AS ir2l'J-2S59 11 27 2;}. l -29 15 27.5 11.71 0.0231 93.0 0.151 Sy 2 1.0 .K 0.1 NCIC 1507 12 ;{5 ;»i.7 -39 51 31.5 12.92 0.0132 52.8 0.250 Sy 2 \.7 s 1.3 IC 3(i:}9 12 -10 52.8 -3(i 15 22.1 13.00 0.0125 50.0 0.2-12 Sy 2 1.2 .\ 1.2 IRAS i;J0r)9-2i()7 13 OS 12.0 -21 22 58.0 10.00 0.0111 50.1 0.273 Sy 2 .\ NCC' 5135 13 25 11.0 -29 50 00.2 12.88 0.0132 52.9 0.250 Sy 2 2.0 X 1.8 lUAS 1H:{25-592() 18 30 58.3 -59 21 08.2 13.20 0.0192 70.8 0.372 Sy 2 X 1(" 1995 20 19 59.0 -52 37 18.0 11.28 0.0103 05.2 0.310 Sy 2 1.2 X 0.8 IC 50G:i 20 52 02.1 -57 01 00.7 12.89 0.0113 •15.2 0.219 Sy 2 2.1 X 1.1 Nc;c 7i;u) 21 18 19.5 -31 57 0 1.0 12.98 0.0159 03.0 0.308 Sy 2 1.5 X 1.1 NCC 72i:{ 22 09 l(j.2 - 17 10 00.2 11.01 0.0050 20.0 0.097 S.V 1 3.1 X 2.8 "'.i cm iculio |)c>si(ioii.s; i-xci'))! N(l(' f)!:},*) wliich is lU cm 'Taki'ii frum ilii- Nivl) dalahasc 'lakcii from dc (Jriji) cl al, l!)!)2, oxcopt NtiC 71:U) wliicli is taki'ii frimi viiii ilvii Uri>vk lil'.U ami N(K' r>i:jr) wliicli is taken Irom i'liillips cl al. ''Uascil on a iliibhli- conslani of //„ = 75 km .s~' Mpc"' ''ryi)iciilly ScyfiMls an- typcil 28 1 I 1 1 I 1 1 1 I 1 1 1 1 I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 O NGC 3393 • IRAS 1121-2806 IRAS 1124-2859 ^ NGC 4507 IC 3639 26 — X IRAS 1305-2407 — K NGC 5135 • IRAS 1832-5926 * IC 4995 • IC 5063 # NGC 7130 # NGC 7213 24 — - • - X 22 ic 20 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 > 1 1 10 11 12 13 14 log[LFIR /Lsun] Figure 4.1 FIR-radio correlation For sample. .\ote how the gala.xies lie on the "ra dio-bright" side of the correlation. IRAS 1121-2906 has only upper limits for its FIR luminosity. 108 Seyfert galaxy in the southern hemisphere, has Lj.tghz = W Hz"' (Bolton Shimmins 1973). -Approximately 40% of the galaxies lie on the "radio-bright" (.see Fig. 4.1) side of the well-known FIR/radio correlation (Helou. Soifer. ic Rowan-Robinson 1985) and were therefore expected to show composite AGN/star formation characteristics suitable for a study of this kind. The galaxies in the sample have been observed as part of various optical emission-line surveys (Phillips. Charles. (S: Baldwin 1983: van den Broek et al. 1991: de Grijp et al. 1992). In Table 4.1 we list some of the optical properties of the sample galaxies. 4.4 The Observations 4.4.1 Radio Observations Radio observations were made with the 6 km Compact .\rray (.\TC.\) on 1994 .luly 24 and 25. at a wavelength A13 cm and on 1995 .November 19-22. at A6 and A.{ cm. Each source was observed for typically six 20-min periods spread over 12 h. in order to provide good coverage of the u-v (Fourier-transform) plane. The total on-source integration times were between 1.29 and 5.6fj h. Unresolved phase calibrators were observed before and after every target observation. Data were taken in two independent linear polarizations at frequencies centered on 2.378 and 2.496 GHz with a bandwidth 128 MHz for the 13 cm observations. The 6 and 3 cm observations were made simultaneously, with frequencies renternd on 4.79S and GHz with a bandwidth of 128 .VlHz. Flu.x densities were calibrated assuming a flux density for the standard .-VTC.-V calibrator PKS B19-34-638 of 13. 5.8. and 2.8 .ly at 13. 6. and 3 cm. respectively. .-VIPS was used to perform the standard calibration and data reduction. Images were made and CLE.\.\ed using the .-VIPS routine .MX. and then corrected for primary beam effects. The rms flux densities in source-free areas of the CLE.-V.\ed images ranged from 50 to 720 /i.Iy beam"'. Table 4.2 lists the clean beam dimensions and rms noise levels, for each source observed at all three wavelengths. In the case of brighter sources (>100 m.Jy) we used "self- calibration" to provide additional phase corrections. The resulting dynamic range achieved in most of the images was typically 50-100 at 13 cm and approximately 20-30 at 3 cm. although 109 Table 4.2 Clean beam parameters and rms noise .\ame wavelength Beam rms noise (cm) (arcsec x arcsec) (m.iy/beam) .\GC 3393 13 7.22 -v 3.95 0.09 6 4.07 1.91 0.10 3 2.46 X 1.04 0.19 IR.AS 11215-2806 13 8.00 X 3.42 0.14 6 :{.85 X 1.78 O.ll 3 2.12 X 0.99 0.19 IR.\S 11249-2859 13 7.85 X 3.17 0.04 6 3..58 X 1.86 0.11 3 2.00 X 1.01 0.19 .\'GC 4507 13 0.24 X 3.93 0.17 6 2.58 X 1.98 O.IO 3 1.76 X 1.01 0.16 IC 3639 13 5.33 X 3.86 0.05 6 2.78 X 2.02 0.07 3 1..55 X 1.15 0.13 IR.\S 13059-2407 13 7.95 .X 3.74 0.22 6 4.35 X 1.93 0.27 3 2.38 X 1.08 0.20 .\GC 5135 13 6.83 X 3.72 0.09 IR.\S 18325-5926 13 4.75 X 3.28 0.07 6 2.35 X 1.55 0.08 3 1.30 X 0.87 0.10 IC 4995 13 5.01 X 3.16 0.07 f) 2.44 X 1.80 0.13 3 1.34 X 0.98 0.18 IC 5063 13 4.23 X 3.57 0.72 6 2.32 X 1.74 0.51 3 1.31 X 0.97 0.16 .\GC 7130 13 6.19 X 3.51 0.15 6 3.17 X 1.86 0.10 3 1.71 X l.Ol 0.11 .\GC 7213 13 4.93 X 3.78 0.12 6 2.27 X 1.94 0.15 3 1.26 X 1.08 0.16 110 in most cases the images were limited by thermal noise ratio rather than dynamic range. Total fluxes for the sources were calculated by blanking all data below twice the rms noise in the image in a region close to the source. Because these observations are optimized for studying the nuclear region, they are relatively insensitive to extended emission. Thus our images are primarily of the nuclear region and will therefore not show any smooth extended structure on scales of tens of arcsec. and consequently the total radio flu.xes quoted may be an underestimate. Table -L.} lists the fluxes of the sources at each of the three wavelengths and other radio properties. The images have resolutions in the range 1-8". The gala.xy .\'GC ol-io was observed solely at 13 cm. because published 6 and 3 cm V'L.A. images are available for this object (L'lvestad &: Wilson 1989). Spectral indices were calculated, in the core regions, for 13 and 6 cm, 13 and 3 cm. and 3 and 6 cm (a(6,13). a(3.13). and a(3.6) are listed in Table 4.3). by restoring the higher frequency images with the same beam and cell-size as used at the lower frequencies. To calculate the spectral index accurately, we measured the flux den.sity over similar regions for each source, at both frequencies. 4.4.2 Infrared and Optical Observations .\'IR spectroscopy and imaging were performed at the 3.9 m .-\.\T. in 1994 .July and October. The .July observations resulted in only 2 h of useful non-photometric K-band imaging data horause of bad weather. Direct imaging was done using the HgCdTe 128 x 128 array detector of the Infrared Imaging Spectrograph (IRIS) and a field of view of lf3 x lf3. .-Vlthough we obtained K-band images for IR.-VS 18325-5926. IC 4995. IC 5063 and .\GC 7213. only .\GC 7130 shows interesting structure (see §4.6.2). •Vttempts to obtain long-slit. .N'lR spectra during this run were unsuccessful because of bad weather. During part of the .\.\T Director's discretionary time on 1994 October 21-23. non- photometric conditions allowed us to obtain a series of long-slit. .\IR spectra for four gala.xies: IR.\S 18325-5926. IC 4995. IC 5063 and .\GC 7213. .Vt f/15 the image scale was O'.'ei pi.xel"' and the spectral resolution was ~300. The telescope was programmed to step the slit across the gala.xy in 0^'6 increments and integration times of 180-220 s were obtained at each position. In Table 1.3 Hadio Pro|H'rl'u'.s Name Si;t, m " S".()(/»! S" Morpliology'' r»((j,l3)'- a (3,13)'' n(3,())' (m.Iy) (mJy) (in.ly) W Hz-' NC:C 3393 •10.9±1.1 20.5±0.1 13.3±I.O 22.11 L -0.9l±0.03 -0.79±0.02 -0.()2±0.09 ll^AS H2ir)-2K0(j 33.0±0.7 20.0±0.l rj.o±o.K 22.01 C -0.72±0.03 -0.07±0.02 -0.75±0,07 IHAS ll2l9-2Hr)9 33.1±0.7 22.3±0.5 H.Sil.O 22..52 I -0.()2±0.0l -0.73±0.03 -0.79±0.12 N(iC 1507 23.1±2.0 lO.HiO.l H.S±(),(j 21.H7 1. -0.S1±0.09 -0.59±0.05 -0.31±0.11 IC 3()39 •12.0±0.r) 1H.2±0.3 9.5±0.0 22.07 1) -0..S8±0.03 -0.92±0.01 -1.01±0.0(j IHAS 13()r)9-2l()7 31(j±2 215±1 10S±1 23.10 (• -0.7I±0.01 -0.93±0.01 -1.17±0.02 NdC r)i3r, H2±2 22.55 1) IHAS lH32r)-o92() 123±1 (i7.2±0.1 3r..r)±o.3 22.91 (' -O.HHiO.Ol -0.97±0.01 -1.05±0.02 IC 1995 H.2±0.r) 5.2±0.5 3.i±0.7 21.70 (' -0.01 ±0.08 -0.(i9±0.05 -0.71±0.10 IC 5()()3 717±12 389±r) 211±10 23.21 l.+I) -0.90±0.02 -0.9S±0.01 -1.05±0.05 N'CC 7130 97.7±0.7 r)9.0±0.S 3(I.S±0.1 22.05 L -0.07±0.0l -0.8l±0.01 -0.98±0.0l NCC 7213 l(j8±l 207± 1 17(j±l 21.SH (' +0.30±0.0i +0.02±0.01 -0.2()±0.02 "I'lxtciulfil -f coif flux fi'Din m,ANl\('il imiiKc = l.iiii'ivr Clivsti, 1) — l)ilf\ist- ("livr>!i, (' = ('oiii|>;ul ('last. 'Spcciial index for l.'i and 0 cm ''.Spcctlal index lor l.'i and .'i cm ''Spe< tral index for G and H cm 112 this way. a spectral cube was built up giving full two-cliniensional spectral information across the inner dO" of each galaxy. Two sets of scans were made, one in each of the H (Al.fio //m) and K (A2.2 fim) NIR bands. Infrared emission was detected from the Br7. H2 v=l-0 S(l) 2.121 /mi and [Fell] AL.644 /ini lines for all four galaxies observed. Representative K-band spectra are presented in §4.6.2. We obtained long-slit optical spectroscopy at both the .-V.A.T and CTIO of four gala.xies. to search for extended emission associated with their radio and .\TR structure in their nuclear regions and inner disks. The spectroscopic optical observations were made at intermediate dispersion at the .-VAT in 199-5 .\ugust and low dispersion at the 1.5 m CTIO telescope in 1995 .-Vpril. The .\.\T observations were made using the RGO spectrograph, in combination with the TEK Ik CCD detector (1024 x 1024) at f/S with the 25 cm camera and a grating with 270 grooves mm"' in first order to produce a spectral resolution of pi.xel"' over a wave length range of 4170-7625 The spatial plate scale was 0'.'77 pi.xel"'. The slit wa.s set to a width of 1'.'66 for gala.xy observations, which was comparable to the atmospheric seeing. Total integration times at each position were 1200-1''^00 s. These observations were made during photometric conditions for only hours on the first night. The standard star LTT 7379 (Stone Baldwin 19)^3) was observed at appro.ximately the same airnuuss Jis the gala.xies. with a wide slit to collect all the light, for purposes of (lux calibration. The rest of the night, and the next night were wiped out because of poor weather. Two fully flu.x-calibrated data sets were obtained for IC 3639 and .\GC 5135 for various slit orientations and positions. CTIO spectra were taken for two galaxies (.\'GC 3393 and IR.\S 11249-2859) in 1995 .-Vpril with the 1.5 m telescope, in combination with the GEC 10 CCD. at f/7.5. The grating used provided a very low spectral resolution of .S.5 .-V pi.xel"' and a wavelength coverage of 4325- 6750 .-V. The plate scale was 1'.'9 pi.xel"'. The slit width was 3". which was larger than the atmospheric seeing (typically 1'.'2). The integrations were 900 s in length. The spectra were obtained before we knew what the radio structure of our sample was. and so all the slit positions were taken along position angle (P.-V) 90° (X through E). For both the CTIO and .-V.\T spectra ii;} we made no correction for telluric absorption. The fluxes of the emission-lines were corrected for internal reddening using the observed Balmer decrement and assuming a standard interstellar reddening curve (Savage i; Mathis 1979). The relative line emissivities for Case B recombination were given by Osterbrock (1989). We u.sed the SPLOT task in IR.-VF to measure and de-blend the optical features in our spectra. VV'e extracted spatial information of the emission-line gas from the various slit positions, by binning the (lux into increments for the .\AT spectra. For the photometric data, the line flux ratios are accurate to the ±10-15% level in the brighter nuclear regions, decreasing to ±20-30% far from the nucleus. Table 4.4 presents a complete summary of the observations made on each galaxy at radio. .\IR. and optical wavelengths. Table 4.4 Observations made on the sample .Name Imaged in radio? Optical long-slit .XIR long-slit -XIR imaging? spectra? spectra? N'GC 3393 yes yes no no IR.\S 11215-2806 yes no no no [R.\S 11249-28.59 yes yes no no .\GC 4507 yes no no no IC 3639 yes yes no no IR.VS 13059-2407 yes no no no XGC 5135 ye& yes no no IR.-VS 18325-592G yes no yes yes [C 4995 yes no yes yes IC 5063 yes no yes yes XGC 7130 yes no no yes XGC 7213 yes no yes yes 4.5 The Radio and FIR Properties In this section we present our radio observations. In §4..5.l. we provide quantitative and qualitative definitions of the L-. D- and C-class radio morphologies, along with the presentation of the full-resolution A13. 6. and 3 cm contour maps, which are also shown in Figures 4.2 and 4.3. In §4.0.2 we pre.scnt and di.scuss the radio spectral indices derived from our radio data. 114 and in §4.5.3. we discuss the FIR/radio correlation and FIR spectral indices, and their role in distinguishing .\G.\ from starburst activity. 4.5.1 Radio Morphology The flux densities, morphology, and radio spectral indices derived from the radio images are presented in Table 4..}. 50% of the gala.xies observed at 13 cm show e.xtended emission, and ~ show extended emission at both 6 and 3 cm. We find a wide variation in radio spectral indices. Free-free absorption is likely to be important in the cores of two systems: NGC 4507 and N'GC 7213. and in the outer radio components of IR 11249-2859 (see §4.5.2). There is also a wide variation in the radio morphologies of the gala.Kies, which fall into four main classes (e.g. Wilson 1988). The physical significance of these classes is supported by the fact that different optical emission line ratios are associated with the different morphological radio types and that there appears to be some correlation with IR.\S color ratios (see §4.5.3): 1. D—class diffu.se radio morphology which is characterized by a low surface brightness component surrounding a compact core. The low surface brightness emission is generally extended on scales > 5". D-class sources have been linked with starburst activity in that the brightness temperature and spectral index of the radio emission are consistent with a combination of supernova remnants, diffuse synchrotron emission from supernova-accelerated cosmic-ray electrons, and H 11 regions (e.g. L'lvestad 1982: Condon et al. 1982). 2. L-class linear structure which straddles a compact core believed to be powered primarily by an outflow from an .-VGX. The outflow may, however, also arise from a circumnuclear starburst (e.g. Baum et al. 1993). W'e define the radio emission as "linear" if the a.xial ratio of the linear structure is > 3. L-class radio sources often have optical emission line ratios consistent with high ionization, high-velocity gas associated with the narrow-line region (.\'LR) (Haniff. Wilson, i; Ward 1988: Wilson. Ward, i.: Haniff 1988: Whittle et al. 1988). and as such are intimately connected with an .-VG.V. Whittle et al. (1988) studied the velocity structure of the [Oni]A5007 emission line in a sample of L-class sources, and found that in some cases there was evidence for an outflow of ionized gas. suggesting that the radio emi.ssion of some L-class .sources may be due to an outflow of plasma from an .\G.\'. 115 Figure 4.2 13 cm radio - optical overlays. The optical images were obtained from the Digitized Sky Survey (.1 plates). The numbers at the top of the overlays are thousands of counts, representing the stretch for the grey-scale. .-Vt the lower right is an inset showing more clearly the radio structure. The contours are at (-2.5. 2.5. 4. 8. 16. .32. 64. 128. 256. 512. 1024) times the rms noise (see Table 4.2 for rms values) measured in a source-free area of the map. 116 Figure -L2 (Continued) •i. C-cIass compact sources, which are either unresolved or only marginally resolved on the arc.scc scale by our present observations (see Table The D and L classifications have to be used with care, as they refer to the dominant soiirrp nf radio fliix wlurh ran rover a wide range of scales (arc seconds to arc minutes). For e.xaniple. a faint L-class source may be contained in D-class emission, as in N'GC 1068. where a collimated structure is seen inside a starburst disk (e.g. Wilson L L'lvestad 1982b: Wynn- W'illianis. Becklin. ic Scoville 1985). Other ambiguities may al.so arise. .-V D-class source seen edge-on might appear as an L-class source. .V linear radio structure could be the result of an outflow from an or circumnuclear starburst. but could also be emission from a bar or an edge-on spiral arm. These ambiguities can sometimes be avoided by comparing the optical and radio emission. For e.xample. radio emission from a spiral arm or a bar will be coincident with the optical structure. -As shown in Table 4.3. our .sample contains two D-class. four L-class. one is a hybrid L-l-D- class (see below), and five C-class sources. IC 3639 and .\GC 5135. the only D-class sources 117 1 r NCC3U1 0 If <7M.OOOUKZ .^ MSajXMMHZ Nac3M3 " G . 0- 0 0 24-0 23J 73Ji 22^ 22^ 3*^ aj 2X0 22J mCHT A5Cf NSON (J3000) ACHT ASCEMSKM (J2000J iftASim MASf121 IS3a.OOOMKZ 47UMUHZ 0 a ^ ' ,0 3 s (f tf24M.O 013 0X0RICKT ASCCKSION02^ (J20001020 03J fUCKT03JI ASCCNSIONCJJ (JMOOl9Z0 IRAS1124 tRAS1'2« ISa.OOO MHZ t\- 479a.00Qt4KZ j 0 < 20 0 I 30 - r • - 0 1 ° » '•• V.' •' 24M RICHT2Xi ASCZHSOH aj (JSOO) 22J 34J RIGHT2U ASCENSION (J300Q) 2LS Figure -1.3 6 aiici -i cm radio maps. The contours are at (-2.5. 2.5. -t. S. 16. .{2. 6-1. 128. 2-56. 512. 102-1) times the rms noise (sec Table -1.2 for rms values) measured in a source-frce area of the map. IIS 47W.OQOMMZHCC4507 23J- 4 ' 0. 401- 50 j— i t23S3l^ 37 j 370 3SJ 3i 0 \2iS39J0 JTJ fa& G f t "l- 12 40 54J S4.0 UJ 93.0 ilS S2.0 t2 O 94J S-*M S3J ntQKT ASCCNSIOKS2J (AOOOlSZO St.S RtC»ir A5CSNS0N (J2000) 4r- •* itusixsf I WtMMHZlAAStaS•• ' • " 47VIJM0 MHZ ! " «K ' iJ ' -1 •; . i iS 0 - .,U ' If' 42J 42.0 4tJ 4t.O 4ZS 4X0 41J rnCKT ASCeXSiON (J3000t ntCKT ASCINSION (JTOOO) Figure 4.3 (Continued) 119 ^—tt (fustcn » '',4mi)00MKZ.aasitu HMJOOOUHZ I« 00 - « « !> 15— IS 37 00 3691 m 57 msb m s7 right a5cen90m (j3000) acht asccnskin (j?000| -| T : I I • * : TIT •sa jTwpi «*«l* •hc*9«5 tc4995 m3s 00q mhz 47«t.oooi4hz * ^ • • % J j_ t o t . i»c^a 00^ t«60.0 s94 5«0 u3 u.0 $7^ s7.0 tcnot4 00 j ttm.o its si.o st j 57^ sta mcht ASCENSION ic90u 47ssjooo uhz t • i B 03 02 01 03 03 ot wght asccmsiomu3000) mcht asccmsion(j30e0) Figure 4.3 (Contituied) 120 NGCTtU HCCTiaO ' 4799.000 MHZ - SS38U)00MHZ q o> 0 0 0 j: L- 18.0 Z14«21.0 20.S 20J} 19J 19.0 18 S 21 a 21.0 20.5 30.0 19^ t9J) tas RIGHT ASCEHStON (J2000> BIGHT ASCCHSIOW (J2000; n— [ I r f I I f4GC72l3 NCC72t3 8638.000 MHZ *• 4798.000 MHZ 5 1000 - O C5D O o 22Q91t.O 174 17.0 16.$ 16.0 15.S 15.0 22 0918.0 173 17J) 18-5 180 155 HO mCHT ASCCHSION (J2000) ^ RIGHT ASCCt^N(J300a| Figure 4.3 (Continued) 121 in our sample, are nearly face-on gala.vies and their D classification was straightforward and not hampered by any orientation effects. On the other hand, care must be taken in the interpretation of the L-class objects in the sample. In one case. \GC 7130. the elongated radio structure may be due to the e.\istence of an asymmetric stellar bar in the galaxy. This source is known to contain vigorous circumnuclear star-formation (Shields Filippenko 1990 - see §4.6.2). and the long-dimension of its Al.} cm radio structure is oriented along a position angle (P.-V) 0° which is close to that of an inner bar. We therefore attribute the radio emission in this gala,\y to either shocks or star formation within the bar. However, radio emission from the other L-class galaxies probably represents an outflow from the nucleus. The P.\ of the linear structure seen in .\GC 3393 at 13 cm is 23° and is not coincident with spiral arms or a bar. The asymmetric double radio source seen at 3 cm is also seen in recent \T^.\ 6 cm observations at higher spatial resolution (R. Morganti. private communication). .\GC -1507. while not highly collimated. meets our criterion for an L-class. although the knots seen at 6 cm to the north and the south could be due to a "hot spot" ring, extra-nuclear H II regions, or S.\ near the nucleus, and not a nuclear outflow. As we discuss in §4.5.2. the radio spectral index of the nucleus of NGC 4507 is flat, suggesting that there is significant free-free absorption. The position angle of the linear structure seen at 13 cm and (i cm in IRAS I r249-2f^59 is 125° and is at an angle of 25° with the major axis of the gala.xy. It is therefore unlikely that the emission is from H II regions at the end of an edge on spiral arm or from a bar. 4.5.2 Radio Spectral Indices There are four main effects which are likely to influence the radio spectral index of Seyfert and starburst galaxies: I. the synchrotron emission from the interaction of cosmic rays with the interstellar mag netic field. The cosmic rays are generated and re-accelerated by supernovae and supernova remnants, but have a lifetime which extends their range beyond the supernova remnants them 122 selves. This synchrotron radiation is expected to have a spectral index of ~ -0.7. The emission generally dominates the radio power of starburst gala.xies. and is presumably responsible for the tight FlR/radio correlation obeyed by these galaxies: 2. the synchrotron emission from relativistic particles generated by the AGN. This radiation could, in principle, cover a wide range of spectral indices. For e.xample. the cores of radio-loud radio gala.xies and quasars typically have a flat-spectrum core (because of synchrotron .self- absorption) but steep-spectrum extended radio-lobes (because of the cooling of high-energy electrons). However, most Seyfert gala.xies have .AGX core spectral indices in the region of -0.7. Condon et al. (1991) argue that, amongst infrared-luminous galaxies, even those that are flat-spectrum do not have a sufficiently high brightness temperature for synchrotron self- absorption to be important; thermal emission from H H regions. Compact H H regions in our gala.xy generate free- free emission from hot electrons. Most are optically thin, giving a flat spectrum, but most compact H 11 regions are optically thick at centimeter wavelengths, giving a spectral index ~2. However. Condon et al. have demonstrated that, at centimeter wavelengths, the integrated flux of such regions is generally small compared to the synchrotron emission, so that star-formation regions in starburst and Seyfert gala.xies have a typical synchrotron spcctrum: -1. the radio emission from ultra-luminous infrared gala.xies is optically thick to free-free absorption, so that the typical synchrotron spectrum of these galaxies is flattened at low frequencies (Condon et al. 1991). The combined result of these effects in Seyfert and starburst galaxies is to produce a typical radio spectral index of -0.7 with a flattening at low frequencies in some starburst sources because of free-free absorption. In our sample of galaxies, e.xcluding XGC 4-507 and .\'GC 72l:i. the spectral indices of the cores lie in the range -0.57 to -1.05. which is consistent with a dominant non-thermal synchrotron process, although the cosmic-ray electrons may be associated either with star formation, or else with an .\G.\ core. .\GC 1-507 and .\GC 721.'3 differ from the other Seyferts in our sample, having low spectral indices (o(;}.6) = -O.iM and -0.25. respectively). Other Seyfert gala.xies are known to contain I2:i flat-spectnirn radio sources (Ulvestaci Wilson 1989). These authors suggested that the flat spectral indices they derived for some of their gala.xies could arise from synchrotron self- absorption or free-free absorption from the ionized gas. Given that .\GC 721.3 is a Seyfert 1 gala.xy. it is relevant that .Antonucci Barvainis (1988) reported that 8 out of 17 RQQs and Seyfert I gala.\ies imaged at 20. 6 and 2 cm had flat spectral indices, arising from a flat or rising component which becomes dominant by 2 cm. This 2 cm "e.Kcess" could be due to free-free emission either from astarburst (requiring a brightness temperature <10' K) or from ~10'' K nuclear free-free emission (.A.ntonucci Barvainis 1988). The core of XGC 7213 is marginally resolved by our 1" beam at -i cm. .Assuming a core size equal to our synthesized beam, the brightness temperature of the core in this source is about liOOO K. which is too low for synchrotron self-absorption, suggesting that free-free absorption is responsible for the low spectral inde.\. although free-free emission is also plausible. The free- free absorption may arise from a vigorous starburst present in the nuclear region of this gala.xy. or. perhaps, from a high column density narrow-line region. Interestingly. .VGC 7213 contains a ring of H II regions 20-40" away from the nucleus (Evans et al. 1996). which, at the distance of this gala.KV (see Table 4.L). corresponds to a ring 2-3 kpc in radius. Yet. our radio observations reveal that in the inner 1" (or inner 100 pc) star-formation is an important contribution to the radio emission. Radio observations of much higher resolution and dynamic range are important for this source in order to understand the morphology of the radio emission. Unfortunately, the core of .\'GC 4-507 is unresolved at 3 cm so we cannot calculate the brightness temperature of the emission for that source, and as a consequence we cannot rule out synchrotron self-absorption as the cause for the flat spectral inde.K. We have obtained radio spectral indices of the knots in the linear structure imaged at 13 and 6 cm. in IR.\S H249-28o9. The spectral inde.K. Q( 13.6). of the first knot to the south-east (see Figures 4.2 and 4.3) is -0.88. consistent with synchrotron emission, whereas the first two knots to the north-west are -0.23 and -0..34. respectively, which is consistent with free-free absorption. 124 4.5.3 FIR/Radio Correlatioa and FIR Spectral Indices The FIR/radio correlation is thought to arise from a common dependence of the FIR and radio emission on the formation of massive stars (e.g. Lisenfeld. Volk. ic .Ku 1996). and it is generally believed to hold up to high FIR and radio luminosities. Gala.xies. or regions within a gala.Ky. that adhere to the FIR/radio correlation are likely to be regions of massive star-forniation. Our sample contains a mi.x of gala.xies that fell on. or were radio-bright with respect to. the FIR correlation, as illustrated by Figure 4.1. which is based on our 13 cm flu.x densities (see Table 4.3) and published IR.A.S flu.xes (see Table 4.5). .-V comparison between the FIR flu.K and the radio flux is a useful indicator of the dominant FIR and radio emission mechanisms. For example. Helou et al. (198-5) define a parameter q. which is defined as: q s log[(FIR/3.75 x 10'^ Hz)/Si,,9GHz] (0 where FIR = 1.26 x I0~''(2.58SfiOfx+Sioo^). (2) It has been shown that q is an indicator of the relative importance of starburst to .\G.\- dominated activity in the nuclear heating of dust in galaxies. For example, the mean value of q for the IR.A.S Bright Gala.xy Sample starbursts is 2.34 ((T,, ~ 0.19) and for AG.Vs q < 2 (.see Condon et al. 1991). We present in Table 4.5 the q-values for our sample, based on IR.\.S FIR fluxes and on our 13 cm (2.437 GFIz) radio flux densities. We use q = log[(FlR/3.75 x 10^- Hz)/S-2..igHz]- similar^ to that of Helou et al. (1985). .\lso presented in Table 4.5 are the published IR.-VS FIR fluxes (de Grijp et al. 1992) and the corresponding FIR spectral indices (cv = d log S/d log f). Dust heated primarily by a luminous buried .-\.G.\ is found in some gala.xies. (i.e.. the so called "warmer" gala.xies) and \'ader et al. (1993) suggested that such sources emit strongly near 60 /an. whereas the UV emission from a pure starburst population is believed to be re-radiated mostly between 60 and 100 f.im. In particular, gala.xies with star-formation tend to have steeper far-infrared spectral indices (rt(25.60) <-1.4 and n(60.100) < -1.1). and 'Note that wo have used a slightly different definition of q from Helou et al. (1985). .\ssuminR a spectr.ol index between 2.1 and 1.-19 GHz of -0.7. the borderline between /VGN' and Starburst would be shifted by O.l-I in dex between the two definitions, which is a small effect Table l..") I'libli.slied IKAS l''liixe.s, comiMiled ([-values and simple cla.ssilicatioii .sclieiiie Name 12//iii 2.5//111 (iO//iii 100//iii n /i /t log <1 ACN/staibiirst" (•J.V) (•).v) (.).v) (12,25) (25,00) (00,100) [FIK/I.,.] a)m|)usile NCC <0.25 0.71 2.3S 3.91 < -1.12 - 1.38 -0.99 10.07 1.92 IK AS iril.vihog <0.11 0.31 0.59 <1.00 <0.18 -0.71 > -1.03 <9.15 <1.11 IKAS ir2l9-28r,9 0.11 o.;m 0.(j0 0.91 -1..51 -0.05 -0.88 9.92 1.10 NCC 1507 O.Ki 1.12 •1.52 5.10 -1.51 -1.32 -0.35 10.10 2.39 X IC 3(j;{9 0.(i(i 2.32 7.OS 10.77 -1.71 -1.27 -0.82 10.33 2.37 .X 11{AS i:}059-2107 <0.25 0.71 l.'ll 1.75 < -1.12 -0.78 -0.11 9.82 0.72 NCC r)i:{r) O.Gi 2.10 10.91 28.(il -1.81 -2.23 -1.03 10.92 2.31 .X IKAS l.s:{25-592() 0.()0 1.39 3.17 1.09 -1.11 -0.<)1 -0..50 10.19 1.53 IC 1995 <0.25 0.3() 0.90 1.28 < -0.50 -1.05 -0.09 9.77 2.17 X? IC 50(i;{ 1.17 ;{.S7 5.89 1.2(i -1.03 -0.18 0.03 10.19 0.91 N(iC 7i:u) 0.59 2.12 Ki.lH 25.57 -1.75 -2.31 -0.80 11.01 2.37 X N(;c 72i;{ 0.()U 0.71 2.,5{) 8.(j;j -0.22 -1.12 -2.38 9.52 1.13 X "iliii'slioii iiiivrk, for 1(" lOOr), dciioli's (liiit \vf infiT a loinpo^iu- sv!>li'in IVoiii ils IokIi i|-valuf, .sec Ic.xl 126 those systems dominated by an AGN have a relatively fiat FIR spectrum. Of the 5 "compact" (C) radio sources in our sample. 4 (IRAS 11215-2806. IR.AS i:}0o9- 2407. [R.\S 18320-5926. i.: XGC 72UJ) have FIR flux ratios and q-values that support AGN- dominated FIR and radio emission. These systems probably have very little nuclear/circumnuclear star-formation. The one exception to this is IC 4995. This galaxy shows unresolved structure in the ."3 and 6 cm radio maps and yet has a q-value suggestive of a star formation-dominated gala.xy. Spectra, which will be presented in Paper II. show little evidence for significant powerful large-scale star formation in this gala.xy which argues against the FIR flux coming from star formation in the outer disk. Furthermore, the point-like nature of the K-band image of this gala.xy (not shown) also argues .-VG.-VI.N'ST disk-wide star formation contributing to the large FIR flux. We therefore conclude that circumstantial evidence points to IC 4995 containing a very compact starburst region which is unresolved by our present observations, .\lternatively. IC 4995 may turn out to be the e.xception to the rule discovered by Helou et al. (that high q values are indicative of starburst-dominated systems). For this reason it would be highly advantageous to obtain higher resolution optical. IR and radio observations of this galaxy. In Figure 4.4. we present a plot of the IR.-VS S(100)/S(12) flux ratio versus S(60)/S(25) ratio to create a color-color diagram for the sample. This combination of 1R.\.S flux ratios has recently been found to be well correlated in samples of Seyfert 2 gala.xies (e.g. Hcisler. Lumsden. Bailey 1997) and is being modeled by Dopita et al. (1997). It appears that our galaxies also show a tight correlation in these colors. .A. lower limit for the S(100)/S(12) flux ratio is shown for .N'GC 3393. IR.\S 13059-2407. and IC 4995. For IR.\.S 1121-5-2806. only the S(60) and S(25) band flu.xes are known. Reasonable limits on the S(100)/S(12) flux ratio would imply that it too follows the correlation (see vertical line in Fig. 4.4). Figure 4.4 shows a number of interesting properties for the sample. If the radio morpholo gies of the galaxies are taken into consideration it is clear that the C-class galaxies occupy the lower-left (warmer FIR colors) part of the diagram, whereas the D-class sources (and the one partially resolved C-class source) lie at the cooler end of the correlation. The L-class sources 127 100 -I 1 1 1 1 r Oh S a. 10 S(60 /uni)/S(25 ftm) Figure -l.-t A FIR color-color plot of the IRAS S(100 /mi)/S(r2 /nn) flux ratio versus the S(60 //in)/S(25 /mi) ratio. Solid squares represent the C-cla.ss sources, the open square with the circle inside represents the marginally resolved source (N'GC 72i;{), the crosses represent the L-class sources, and the open circles represent the D-class sources. In the case of IR.-VS 11215-280(5. only the S(GO) and S(25) fluxes are known, and the unconstrained S(100)/S(I2) ratio is denoted by an open square with a vertical line running through it. Interestingly, there is a segregation of the morphological classes in the color-color plot: the compact sources lie in the lower left (warm FIR colors) and the diffuse sources (and the marginally resolved source) lie from the middle to the upper right (cool FIR colors). The L-class sources span the whole range, but the most highlv collimated sources reside in the lower left. 128 span the entire range, although the two most highly collimatecl sources. IC 506.'} and IRAS 11249-2859. lie in the warmer region of the diagram. This morphological segregation, based on IR.\S colors, suggests that the global FIR properties are linked in some way to the radio morphologies. To understand how this might be. we will consider what has been learned recently about galaxies with small (i.e. warm) S(60)/S(2.5) IR.-VS flux ratios. Heisler et al. (1997) have shown that Seyfert 2 galaxies with warmer S(60)/S(25) colors are most likely to show broad permitted lines in their polarized nuclear light (Hidden Broad-line regions-HBLRs). and lower line-of-sight extinction to the nucleus. Within the context of the unified theory of Seyfert gala.xies. Heisler et al. provide evidence that the warmer colors are a result of the viewing orientation of the inner dust torus which is postulated to surround the .-VG.V. A dust torus seen closer to face-on would show warmer IR.-VS colors than those seen more nearly edge-on because the viewer sees more deeply into the inner regions as a result of a higher viewing inclination. Can the same argument be applied to our sample? IC 506;} (L+D) shows the warmest colors in Fig. 4.4 and is already know to be a HBLR gala.xy. and so its warm colors are consistent with the Heisler et al. picture. .\ more challenging problem is to explain why the C-class sources should have warmer IRAS colors. One possibility might be that the compact sources are collimated "jet-like" sources seen end-on. If the jets were emitted roughly orthogonally to the dust torus, then this would provide a natural explanation for the warmer colors since the viewing direction would be favorable for seeing deeper into the dust torus. In at lea.st one of the C-class sources (IR 18.'}2-59) the gala.xy exhibits high e.xcitation optical emission lines (Isasawa et al. 1995) and is suspected, from .\-ray evidence, to contain a Seyfert I nucleus. This would seem to support the view that we can almost view directly the Seyfert I core. Certainly the two galaxies with the coolest colors, SGC 5135 and .\GC 7130 show evidence for large-scale extended star formation and their FIR colors are likely to be dominated by heat sources spread over a large scale. However, we caution that our understanding of correlations like that in Figure 4.4 is still in its infancy, and any possible correlation between warm IR.\S colors and radio morphology may 129 be entirely coincidental. Work such as modeling by Dopita et al. (1997) may help to clarify the nature of such a correlation in terms of optical depth effects in the dusty torus at 12 /an. We note that Rush et al. (1996) found a similar trend in a mid-infrared selected Seyfert galaxy sample and an optically selected Cf.A. Seyfert sample: radio sources that were more compact tended to have warmer S(60)/S(25) colors. However, the resolution of our data is much higher than that of Rush et al. making a meaningful comparison difficult. 4.6 Optical and Near-Infrared Data and its Relation to Radio Structure In §4.6.1 we present the and CTIO optical spectra, and we discuss the correlations of the line ratios with radio structure for IC 3639. N'GC ol3o. .\'GC 3393. and IR.A.S U249-2S59. In §4.6.2 we present the .\.\T .\IR spectra and imaging, and in §4.6.3 we discuss in detail the radio structure of IC 5063 and its relation to published optical imaging and spectroscopic observations. 4.6.1 Optical Spectroscopy of a Subset of the Sample 4.6.1.1 IC 3639 IC 3639 has been observed spectroscopically and has been imaged in the the Ha line (van den Broek et al. 1991: known as gala.xy 24SW in that paper) as part of a study of galaxies with large FIR-to-optical luminosities. This Seyfert 2 gala.xy shows considerable extended Ha emission over much of its compact, high surface-brightness disk. The overall e.xtent of the disk is only 3 kpc. nearby companion shows a ring of H H regions (van den Broek et al. 1991: gala.xy 24.\E). IC 3639 has been shown, from optical spectropolarimetry. to contain a hidden broad line region (Heisler et al. 1997). Our 13 cm contour map of this gala.xy shows a very rich structure. Surrounding the central core is a diffuse radio structure with several small unresolved knots of emission. son\e of which are coincident with Ha knots, particularly in the northern half of the galaxy. This gala.xy is clearly a D-class radio source. The overall envelope of the radio emission down to the level of 0.16 m.ly beam"' follows closely the boundaries of the optical gala.xy. suggesting that the radio emission is intimately related to disk processes. 130 i.e.. star-formation and supernova events. Our spectroscopic observations from the A.-VT reveal highly extended [01] A6300 and Ha emission. We were able to map spatially the optical emis-sion-Iines in and near the Seyfert nucleus. Representative spectra are presented in Figure 4.5a.b.c. In Figure 4.6. we plot the line ratios obtained for IC 3639 on the standard optical diagnostic diagrams (Baldwin. Phillips & Terlevich 1981: Veilleux Osterbrock 1987) [OlII] A.5007/HJ vs [SH] AA673l,6716/Hri. [01] A6300/Ha and [.Nil] A6584/Hci for PA 90°. centered on the nucleus. The line ratios change from .A-GN-dominated emission in the nucleus, to H Il-region-like emission at a radius of ~1.5 kpc. which suggests that circumnuclear star-formation is important. Note the very blue continuum in Figure 4..5c. which suggests a very young stellar population. Tables 4.6. 4.7. and 4.8 contain emission-line fluxes and flux ratios for the detected lines at I'.' t intervals from the center for the three observed slit positions (see below). It is interesting that in two of the line diagnostic diagrams, those involving ratios of [01] A6300 and [SH] to hydrogen (Fig, 4.6b.c), the points representing emission just out side the nucleus encroach significantly into the LI.\ER region. W'e will argue below that this may be a result of enhanced [OI] and [SII] emission caused by weak (150 km s~') shocks in the circumnuclear environment (Dopitai: Sutherland 1995). marked east/west asymmetry exists in both the emission lines and the radio distribution (e.g. in Fig. 4.6a.c. note how the line ratios show this east/west asymmetry). We will discuss this in relation to a simple model of the galaxy below. .\long P.\ 0° (see Fig. 4.7a.c) the line ratios change smoothly from AGN" to H Il-regioti-like emission with increasing distance from the nucleus. .-Vgain we note that the [Ol] emission-line ratios dip significantly into the LI.N'ER region. L'nlike the spectrum taken at PA 90°. there is no noticeable asymmetry on either side of the nucleus. We also obtained a spectrum T'.'o .\ of the nucleus along P.-\. 90° (see Fig. 4.8a.c for BPT and \'0 diagrams). The [01] distribution is remarkably asymmetric, showing strong LIXER- type emission in the east, but H Il-region-like emission in the west. This mimics the behavior at P.-V 90° through the nucleus, except that no AG.X emission is .seen. 131 I.OOE-U 4.00E-15 2.00E-15 5000I 5500 6000 6500 V«v«lingth (ongstrottsf I.00E-I5 B.OOE-I6 6.Q0E-15 5000 5500 6000 6500 n Wavat •ngth Cangviroasi "T"- 5*500 6000 c V.wc I nnq(h (iinqn trn«^ I Figure 4.5 Spectra of IC 3639. P.\ 90°. going from .\GN (a) to LINER (b) to H H-regioii-like (c). .\ote in that in (b) we marginally detcct an M.i absorp>- tion feature. Confirmation of this feature is pending. The vertical scale is F,\ (ergs cin"-^ s~' and the wavelengths are in the rest frame of the galaxy. i;}2 r '' I ' ' ' i • ' • 1 I''' t •' ' 1' '' I'' ' 1''' I r C3aSPA90 s 5 3 o 0 S V Ean o(mjdaut • WMtofoudius n NuOffw* I n,i 1111111 i t «-i' 1 ' I' • • 11. • f_,,. I,,. i,. t4 12 I #423 i^dmiVKo) \ , I 'r \ « (mcn I3 « t v:\ • 5> p « F«« ql AuOntt r • Wmtol nijcttu* f- u nucmus r tr till'. l „i i t_i 11 ,'_i i. 111,1.11 i•. • i.,• i.. • i 14 i 2 t .9 .$ .4 2 g 2 Loc([SllI/Ha) ' 5 \ J of- \ • • Easiefnuewut • Wtsi ol ntjdaus 'I Nixtous \ i I I I -i_ I I I I t I \ 15 .« 5 Uc(rOIl/Ho} Figure 4.6 BPT and V'O diagrams for IC PA 90°. For this figure, and for the following BPT and VO diagrams: (a) is log([.\Il]/Ho) vs log([OIII]/HJ), (b) is log([SII]/Ha) vs log([Oin]/H.J). (c) is log([OI]/Hft vs log([011I]/H.i). and the points proceed radially away from the nucleus .-VPPRO.XIM.\TEL\' from upper right to lower left with I'.'4 increments. Of interest in these figures is the east-west asymmetry. To the east of the nucleus, where there is a rapidly declining radio emission, the optical emission is consistent with a LI.VER and to the west, the emission is more like that of H Il-regions. This effect is seen in all three diagnostics. I'aUlt' •!.() K' riHliiiMiiiig-iDiriHli'd iMnission-liiio ratios, rolativo to 11,^ = 100, at P.A. 5)0° cciilcrcd on nucleus (list a lice" 11/^'' Av'" [0111] [0111] [01] [Nil] [Nil] [sii] [,SII] (arcst'c) (x U)~'^'tMgs/s-Cll\^) (mag) Al95y A5007 A0300 A0518 A0583 M)717 A0731 -19.0 1.;} 1.01 o;{.;{ 215 30.2 23.0 105 80.9 51.0 -18.2 5.0 2.71 101 301 02.1 27.1 117 01.7 55.2 -Ki.K (il.O i.:ii 211 093 50.8 82.1 239 57.2 03.1 -if).! 5.:i 2.01 110 2«0 29.0 30.0 no 00.1 •19.0 -11.0 rj.r) 2.05 107 290 23.3 39.7 121 75.0 50.3 -12.0 :f.() 1.92 99.1 305 20.1 13.8 MO 78.9 50.1 -11.2 2.:f 1.09 127 310 •18.9 103 70.0 03.1 -{).S 0.5 2.71 191 502 31.0 50.1 159 82.5 •19.3 -K.l 2.5 1.55 115 302 31.1 •18.0 150 71.2 •18.1 -7.0 •i.i 1 .,<0 95.3 275 23.5 •17.1 115 (iO.O 50.1 -r).o O.N 1.2.} 90.H •185 29.1 01.0 100 07.1 57.9 - 1.2 13.1 0.S7 235 050 •11.0 09.5 219 50.0 51.9 -2.N 01.0 i.;n 211 093 50.8 82.1 239 57.2 03.1 -1.1 :i72 2.20 259 715 50.5 89.0 210 51.0 59.1 0.0 K(>:i 1 .:i5 330 891 55.1 97.3 238 50.8 09.7 + 1.1 51.1 3.29 201 399 •17.7 88.0 210 01.3 52.1 +2.« 10.9 2.15 os.o 153 17.9 50.2 138 57.0 12.3 "NcK'ilivtt aiiKular tlisiiiiicfs an- to tlif i-.v-st tin: n\nii'>isi, posiiivt; auKular ilist anci's ari' wcM of llic iiiu'lcus, and l).0 indicates I lie nucleus 'lln fluxes niay In- obtained by nuilliplyiuK thesis (luxes hy J.M 'calculated from the lialnier decrement iissuiniuK Case H rec(inil>inati(in 'lahic 1.0 (('on till lied) dislaiuc" Av*' [0111] [0111] [01] [Nil] [Nil] [Sll] [Sll] (arcscc) (.\ iO~''''(«r{!,s/.s-cni^) (luas) Ai<)r)9 Ar)007 A(i:iOt) A(>5.18 A()58:i A()717 A()7:ii + \:2 IK.l 2.19 ()0.2 i;m 2;}.8 •i(j.;} 128 59.1 •12.0 +r).() 7.:i 2.27 8;i.() i;{2 19.() lli.O Ml 07.8 51.0 +7.0 1.9 1.2-1 (j8.() 121 20.9 •I9.;j VM 80.9 57.2 +s..i 2.1 O.OK ;j8.9 90.1 12.5 109 55.8 •11.0 +!).8 •1.8 1.01 af).! 90.0 (i.2 ;{3.9 9(j.2 •17.1 30.2 + 11.2 l().(j 1 .?srj ;uj.i 101 31.8 90.5 •15.5 35.3 + 12.0 (i.O 1.85 11.8 lor) 15.1 ;n.i 98. 53.1 37.0 + 11.0 •l.H 1.91 10.;} 111 29.8 93.7 57.1 •10.1 + ir)..i a.i 1.78 ():i.2 115 21.9 28.-1 98.2 (51.2 •1 1.1 + l(j.H •i.;{ 2.22 ()1.2 2;{a 27.0 97.9 59.5 •17.2 + is.2 2.2 1.91 101 199 22.5 10.9 107 70.1 •10.1 "NvUnlivf iiiiK\itar ilisUimi's an- U) llu- i-ast of llii: mu'luiis, poMtivc aiiKular ili.slanci'.'. an- we.-.! of ilii- iiiiclcu.s, ami 0.0 iiuli('a(c!i clu- luiclcu.s ''Ho fliixc.s may he ohtaiiicd l>,v mull iplyiiiK llic.se (liixt-.s hy 'calciilad'd from (lie iSalmcr (Iccrcmciil itisiimiii)^ (' (lisliiiicc" 11/^'' Av'' [0111] [0111] [oi] [Nil] [Nil] [Sll] [Sll] (arcscc) (x 10~'^'('rg.s/s-ciii'') (mag) A1959 A5007 A0;{00 A05I8 A0583 A0717 AO 731 l.:i 1.22 •19.1 I;{0 9.7 23.8 70.2 •15.7 30.1 -1 1.0 •1.9 1.;») •18.2 122 10.5 20.1 81.7 •18.8 30.0 -12.0 :i.:i 0.91 ;m.o 105 11.1 29.2 87.5 51.7 38.9 -11.2 :i.i 0.8:i :}:{.7 91.0 11.0 29.3 91.7 55.1 37.2 -!).S :{.o 0.72 28.7 8.1.1 19.2 30.0 91.1 51.8 31.5 -S.I •1.5 0.53 28.5 77.1 12.9 30.5 90.1 •17.5 33.0 -7.0 5.9 0.19 ;n.i 81.0 9.1 28.5 85.9 •11.8 31.9 -5.0 10.0 0.91 31.0 82.2 0.1 27.1 83.8 •11.3 29.2 - 1.2 10.0 1.12 :H.I 71.1 0.8 30.9 91.5 •15.1 33.9 -2.S 8.1 1.00 29.1 80.0 9.9 33.7 90.2 19.7 35.7 -1.1 7.9 1.00 20.9 8:}.0 H.2 32.8 90.3 •17.7 35.8 0.0 11.9 1.19 ao.a 72.5 5.8 31.5 100 •18.9 35.0 +1.1 21.1 2.07 28.8 09.7 9.;{ 33.2 105 52.0 38.1 +2.8 2.11 ;}7.I 75.9 30.5 108 50.7 12.3 + 1.2 0.1 1.01 ;{0.9 77.5 33.0 113 57.5 •10.2 +5.0 •1.1 1.28 29.1 07.7 35.7 no 01.2 13.7 +7.0 0.1 1.80 :{o.o 85.:{ 10.2 31.0 107 00.0 •11.2 +8.-I 1.5 1.70 .{•1.1 71.1 25.0 102 05.0 •11.8 +9.S •1.7 2.12 12;{ 30.9 100 05.2 •18.5 + 11.2 5.1 2.;i5 121 30.9 89.3 00.2 10.1 + 12.0 •1.5 2.28 158 32.2 91.7 01.0 50.0 "Ncgalivt' angular (iisfaiicc.s arc to (lie casl, iiuhilivc aiiKiilar di^laiKcs arc west, aiul 0.0 iiulicalcs tlic poiiil 7'.'r> ilircctly iiorlli »! llic iiuclciis ''11(1 lliixcs may lie obtained hy iiiiilliplyiiiK llic.sc flii.xc.s l)y J.M '(iilciilalcit rniiii (he lialiiicr (U'crciiiciil ahstiiiiiiiK CaM' li recomlHiiaiioii 'ral)lo l.H IC rcddciiiiig-corm led cmissioii-liiK' ratios, rolativt- to 11/^ = 100, at I'.A. 0" rt'iiU'ii'd on iim ltMis distain'o" Av'- [Oil!] [0111] [Ol] [Nil] [Nil] [Sll] [Sll] (airscc) (.\ 10~''''i'rg.s/.s-ciir^) (mas) A1959 A50U7 Mum A()5I8 A()58:i A0717 A07:il -1{).(5 1.1 i.7H 05.1 151 35.7 90.1 90.1 •11.7 -1K.2 l.f) 1.11 59.2 192 I8.:i :i5.c 90.0 82.0 •19.;} -Ki.H 2.H 1.51 (il.l ia9 11.1 •27.7 91.9 50.7 •15.8 -IT).! ;j..s 1.().{ 55.0 98..I i;i.;5 29.1 91.7 57.7 38.0 -1 1.0 •1.9 l.HU •15.0 115 9.8 27.7 100 57.0 •11.5 -r2.() •i.(j 1.01 ;M.7 97.8 12.1 iJO.O 101 55.8 •12.7 -11.2 •1.0 1.17 •19.9 81.;j ;w.i 102 51.0 ;}7.8 -<).H s.o 1.9H 81.8 8.7 •29.2 107 52.8 ;}i.7 -K.l 17.0 •2.50 112 20.9 :{7.i \tl •17.1 :}s.i -7.0 20.() •i. i;{ 58.:} 1 10 21.5 •10.;} i;}l 55.8 11.1 -5.(i 25.1 •2..55 KM •122 ;{i.i 50.;} 17;} 58.;} 51.2 - 1.2 59.9 2.10 :K)2 810 50. 80.5 2^28 00.;} 58.1 -2.8 111 1.88 2(i2 7;M 50.8 9;}.7 •2i;} 57.0 59.5 -1.1 ;ii7 2.02 2(i:} 7:i7 51.1 92.1 •211 57.7 59.0 0.0 1K9 2.72 297 818 58.2 89.9 2;}7 51.9 57.3 +1.1 00.3 2.80 19! 008 •19.1 OO.l 188 50.1 •17.9 +2.S 22.5 2.00 00.1 i;59 15.1 ;}9.0 r27 •10.2 35.1 "NvHiUivi' aii)^uliu' ilistiuues iuv to tin- uf tlic iim:K-vi», iuif-nlar tlihiancf.i ari' lunili ol' llii- nucleus, and O.t) indicalcs llic iiucltrus ''ilo fluxes may he ohiaiued liy mulli|il.viii)4 these fluxes hy J.Kf) 'calrulaled from (lie lialmer (lecremeiit JissuiuiuK Case li recomhiiialiou l.iS (Coiitiimcd) (list a lice" 1 Av'" [Olll] [Olll] [01] [Nil] (Nlll [Sll] [.Sll] (arcscr) (.X 10~''Vrgs/s-cm^) (mag) A1950 A5007 A(i300 A0518 A0583 A0717 A0731 + 1.2 17.!) 1.50 21.7 71.9 13.1 31.2 105 •15.1 33.8 +.').() lO.X 1.58 20.0 72.7 0.7 30.9 97.0 •11.1 33.7 +7.0 20.5 1.02 ;{0.2 91.2 9.3 30.1 92.2 11.1 31.0 +S.1 11.1 l.Sl 30.1 108 32.8 31.3 92.1 •17.2 31.0 +9.H K.-l 1.51 101 12.2 28.5 87.1 •15.1 31.7 +11.2 7.;{ l.K) •10.9 118 8.0 2(i.7 K2.0 •13.5 33.3 + 12.(j }).() 1.23 19.9 135 8.5 23.5 75.9 15.2 32.7 + 1 1.0 8.2 l.K) 11.3 131 13.2 21.1 70.5 •13.8 31.0 + l.'j.l •1.9 I.;M ai.o 133 18.8 23.1 77.5 51.3 31.2 + 1().H 2.2 l.OS 17.;} 132 20.(i 81.(i 58.9 38.9 "NepaliVI- angular liisiaiiccs ari' lo llic .soiilli of llic Miicli'iih, poAilivc aiiKiilar ilisiaiicch arc norili of llic iiuilcii.s, aiul ().() iiulicaCch clif niirlciis ''llo lluxi's may In? ol)laiiu'tl l>y nuilti)>lyinK llu'sc Ituxcs l)y 'calciilalcd from llif Halmcr ilccicmciil assuming Case U rc'i'oml>iiiaiioii i;{s C3639PA0 < North ot nudeus nK SogthNudvus of nud«us Lo|((NIIl/Ha) IC3639PA0 imcn < North of nudsus K Soulh of nudtus Nudeus 6 4 -2 LOC((SHJ/HO) C 3639 PA 0 unirn — « North of nudeus K South of nudeus • • Nudmis I t t t •25 15 1 Lo*((OI]/TIn) Figure -L7 BPT and VO diagrams for IC 'Mi'-iO. PA 0°. Again, tiie optical emission pro ceeds. radially away from tlie nucleus, from AG.\ to H Il-region-like. Note that south of the nucleus the emission proceeds to H ll-region-like at a larger relative distance than to the north. The emission to the south dips into the LI.\ER region (at ~o'.'6). but only significantly in the [01] diagnostic. 139 C 3839 PA 90. 7^ N LoCUHUl/Ha) 10 3639 PAgorrN •cnfcn — * East - « W9St lo«(ls»1/ko) IC3539PA90 7rN U<([OII/HA) Figure 4.8 BPT and VO diagrams for IC .3639. P.\ 90°. T'.'o North of the nucleus. This emission is consistent with H Il-regions. Of interest is (c), which mimics the east-west asymmetry seen in Figure -l.Ga.c. To the east, just north of the nucleus, the emission is consistent with a LI.NER. and not until east docs it dip into the H ll-region portion of the graph. 140 The radio emission is highly extended and asymmetric at 13 cm. being stronger and more extensive to the north-west (see Fig. 4.2). The two spectra which probed the east-west character of the gala.xy also reflect an asymmetry. Strong optical emission on the western side of the nucleus, in the region of the high surface-brightness radio emission, shows H Il-region- like spectra, whereas to the east the emission is more characteristic of LI.NER emission, and this occurs on the edge of the rapidly declining radio emission. This behavior is seen in both east-west spectra which cut through the .\G.\' and to the north of the .AG.X. We note that LL\ER characteristics (especially enhanced line ratios of [01] A6300 and [SlI] AA6716.67.31 to Ha) could arise from either a low-powered .\G.\' illuminating ambient narrow-line region gas or from high-velocity shocks (See Dopita Sutherland. 199-5 for discussion). The fact that the spectra appear H Il-region-like over the area of the extended radio emission region but LL\ER-like at the edges of the radio structure, suggest to us that the LINER emission may be from shocked gas resulting from a tipped outflow from a widespread starburst in the central few kpc. The morphology of the radio emission itself is reminiscent of outflows such as those seen in starburst systems like .\[.S2. Unfortunately, because the radio surface brightness of the very extended 13 cm emission is low. it is not possible to determine a spectral index of the plasma, but the radio emission may be from relativistic electrons which have burst out of the disk asymmetrically, perhaps from supernova explosions. In this simple picture, gas from the outflow interacts with ambient ISM matprial to shock and produce a LINER spectrum. If the LI.NER spectrum were interpreted as coming predominantly from shock waves, they would imply supersonic flow velocities of the order of 150-200 km for IC 3639 (Dopita Sutherland 1995). One can test this model further by obtaining higher dispersion spectra to look for mass-motions in the gas in order to identify conclusively a superwind in this galaxy. .-Vn indication that nuclear star-formation may have been episodic in IC 3639 comes from the marginal detection of Balmer absorption (HJ) in a ring at about 7" from the nucleus. Three l-d spectra, at that angular distance from the nucleus, hint at Balmer absorption. However, it is not seen in any of the other extracted I-d spectra. If these features are confirmed (e.g. by deeper spectra in the blue) this might suggest evidence for a post-starburst population sotne 141 1-3 Gyr old. The HJ emission line, with faint absorption superimposed, was not corrected for absorption and we caution that the Av derived from the Balmer decrement may be unreliable. .-Vs a consequence the line flux ratios, in these three spectra, could be in error by ~ . We show this feature in Figure 4.ob. 4.6.1.2 NGC 5135 N'GC .^135 is a barred spiral. The 13 cm radio image (see Fig. 4.2) shows considerable e.xtended structure. Our observations show that, in addition to tiie central extended emission regions, there are knots of emission on a larger scale. In particular, there are two relatively bright knots of emission, one at a position angle of 10°. 18'.'4 (4.7 kpc) distant from the central source, and another at 279°. 13'.'8 (3.5 kpc) away from the center of emission. Comparison with the optical image shows that the strongest emission is located at the center of the bright optical bar (see Fig. 4.2). L Ivestad <51: Wilson (1989) obtained 6 and 20 cm images of N'GC 513-5. providing arcsec resolution with the .\ and B arrays of the VLA. .-Vt both wavelengths they saw an asymmetric structure consisting of a bright core with amorphous, fainter extension to the north-east (D-class source). The e.xtension broke up into several smaller knots giving it a bubble-like appearance. All of this structure lies within the inner contours of our L3 cm map. We obtained optical spectra at the AAT for this galaxy along P.\ 19° and 15°. The former intersects the radio knot to the north-east and the latter follows the diffuse radio emission detected by L'lvestad &; Wilson. Along both position angles, the emission is extended over ~.5". The optical diagnostic diagrams (see Fig. 4.9a.c for representative spectra along P.\ 19° and Fig. 4.10a.c for BPT and \'0 diagrams: also see Table 4.9 for flu.xes and flux ratios) along P.\ 19° reveal that the emission is that of an AGX to the south-west, but changes in the north east (where one of the radio knots resides) from .\G.\' to LI.\"ER to H II region-like. .Mthough not very extended, the optical emission is consistent with star-formation in the circuninuclear region. .-Vgain. this confirms that a D-class source has significant amounts of star-formation. Unfortunately, the observing conditions for NGC 5135 became poor, and it is possible that 142 1 I • I - ik. mccmura ; 5.ooe-i5 1 ,s — 4.00E-15 1 :.ooc-i5 1 1 - ij i 1 i: ; - 2.0flE-I5 'l 1 ii I •{ b i V I.OOE-IS I ii! 1 ! « —'tv'C y,^ iL". J ijtr ' 1 1 1 - y 5COO 5500 6000 Uftvaltngtn (Angstroas) 2.00E T' KW .4 •WW I.50E-I5 l.OOE-15 - li -1 1 i| 5.00E-16 • I W VI 5000 5500 600C 6500 Wtvslfngth langstroasl *cc3i«st*a iv. t* nwflrf a.00£-I6 - 6.00E-16 - 4.00E-I6 5000 5500 6000 6500 tftvtltnglh (tngstr9«tl Figure 4.9 Spectra of XGC 5135. along P.A. 19°. going from (a) to LIXER (b) to H Il-region-like (c). The vertical scale is F.\ (ergs cni~- s~' A~') and the wavelengths are in the rest frame of tiie galaxy. 143 r-Y- -T-r-i-r-r-rr r f'' ' « | | ' 1 ' ' ' I ' ' NGC S13S PA 19 ^ 0 X ME • SW n Nudeus I • • • I • •. I • • • I U(([Hnl/l{a) A NGCStaS PA 19 § 5 l»cr a* h -•« SWme n Nudeus i 1 ' ' t ' t 1 i 1 1 1 t t » i i . i i 1 . • i i i i t i B 6 -4 -? 0 2 U(([Snl/Ha) t I I' r' • r- T~ '[• I T" 1 ' ' ' ' I NCC513S PA 19 oa am 0 >« NE h • SW o Nudeus _i I I \ » I 2 15 c La(((0(l/lla) Figure 4.10 BPT and VO diagrams for NGC 5135. P.\ 10°. Of interest is a nortii-eaiit soutli-west asymmetry. To the north-east (there is a radio knot at 13 cm along this position angle, see Fig. 4.2a) the emission goes from .VG.V to LI.XER to H Il-region-like. whereas to the south-west the emission remains consistent with an .-VG.V. ral)k' 1.9 NdC ."jliir) rtHldoning-rorrcctcd cMni.ssioii-liiic ratios, rclativo to \\,i = 100, at I'.A. 19° iciilcrc'd on muloii.s di.slaiicc" Ufi'' A/ [01II] [01II] [01] [Nil] [Nil] [Sll] [Sll] (arc.scc) (x lO^'Vigs/.s-nii^) (mag) A19.'j9 A.5007 A(i;«)0 A0518 A0583 A0717 AG731 -•2.8 Ml •1.10 2()8 017 i8.:{ 73.1 198 •10.9 30.8 -1.1 2ua 2.()H 20r) 572 i().y 80.2 222 13.9 38.3 0.0 1-10 1.9() 151 •109 22.1 97.9 253 50.9 •11.7 + 1.1 Pif) 1,87 1 10 •108 21.0 110 281 (>1.5 •18.2 +2.8 89.(i 1.8.') IKi 321 27.5 110 287 (i().5 •19.3 + 1.2 ;}7.(j I.7() 85.9 2;{2 29.5 •13.5 111 27.5 21.0 +.').() 2;{.() 2.19 (il.O l(i7 1 1.5 Gl.l 109 •17.0 •10.7 "NcK'ilivc aiiKuiai' tli.staiicc.s air to (he Mjiilli-wchl of llie laiclciis, positive angular (ii.siaiui'.s an- iu>i'tii-t'a.si of ilu! iiiicli'iis, and 0.0 inilii'aic.s ilic nncicns '"11(1 fluxc.s may l>i' ()l)iainf fainter off-nuclear emission was not detected. 4.6.1.3 NGC 3393 .A. CTIO spectrum was obtained for .\'GC ^J393 along a position angle of 90° through the nucleus. Because of the very low dispersion, we were unable to de-blend the Hfv and [Nil] AA6o4.S.r)o84 lines, and so were unable to construct line ratio diagnostics in the nuclear regions. In both the east and west directions. [01] A6300 is e.xtended over ~2" which, at the distance of this gala.xy. is ~oOO pc. The Ha is more extended, covering ~10". or ~2.6 kpc. to the east and ~6". or ~1.6 kpc. to the west, with a similar morphology to that of the 6 and -i cm radio emission. The extracted 1-d spectrum for the nucleus of XGC is presented in Figure 4.1 la. 4.6.1.4 IRAS 11249-2859 spectrum was obtained for IR.A.S ir249-"28.59 along a position angle of 90° through the nucleus, cutting through the radio knot to the east. We show in Figure 4.lib a l-d spectrum of the nucleus of IR.-VS 11249-2859. It is interesting to note that the Hcv is extended over ~6". or 2.7 kpc. and [Ol] AG.'JOO is e.xtended over ~2". or ~1.4 kpc. Proceeding east, away from the uucleus. the [01] emission at Rrst decliues and thou increases in strength ~4" from the center and then disappears again at larger distances. This increase in flux at 4" corresponds to the position of the south-east radio knot. .A. similar increase in line emission at the position of the knot is seen in the (blended) Ha ic [NTl] AA6.548.()584 emission and the [SII] AA673I.67I6 lines. The ratio of [SII] to Ho is 0.;{9. [.\Tl] A6584 to Ha is 0.48. and [OI]/Ha is 0.04. Off the nucleus. H.i and [OIII] A5007 disappear completely, thus we cannot construct the [OIII]/lI J ratio which would better determine if the emission is .\G.\'-like or consistent with a star-formation region. We note that these ratios are also consistent with a low to intermediate velocity shock-wave process (Dopita Sutherland 1995). Given that the optical emission lines are coincident with one of the components of the linear triple radio source (the SE knot), it is clear that follow-up ob.servations of higher 146 i NGC JW.1 PA 90. Nucleus 6.00E-14 - 5.00E-14 4.00E-14 3.00E-14 2.00E-14 1 1.00E-14 5000 5500 6000 6500 Wavelength (angstroiisl IRAS I PA 90. Nucleus 1.75E-15 - 1.50E-15 1.25E-15 1.00E-15 7.50E-16 5.00E-16 2.50E-16 5000 5500 6000 6500 B Wavelength (angsirons) Figure 4.11 CTIO spectra, (a) N'uclear spectrum for NGC along P.-V 90°. (b) Nuclear spectrum for IRAS 11249-2859 along P.\ 90°. The vertical scale is F.\ (ergs cm""^ s"' A~') and the wavelengths are in the rest frame of the gala.xy. 147 sensitivity along a PA of 125° are highly desirable. We speculate that these observations might reveal that the nuclear outflow is causing wide scale shocks in the disk of this galaxy giving rise to the enhanced [Oi] and possibly inducing secondary star-formation, consistent with the flat radio spectral indices of the knots to the .\'\V. 4.6.2 Near-Infrared Spectra and Imaging .-Ml four galaxies observed spectroscopically in the .\"IR. (IR.AS LS.T25-o926. IC 4995. IC 5063. and .\GC 721.3: see Table 2.4) show detectable emission lines of Br"). H-2. and [Fell]. The nuclear spectra are presented in Figure 4.12. L'nlike the other gala.xies observed at .\IR wavelengths. IRAS 18325-592G shows no evidence for an underlying stellar continuum (in particular the CO (2-0) band-head is missing). This is con.sistent with significant obscuration, even at infrared wavelengths, by dust near the nucleus. L'nusually high levels of obscuration in this probable heavily embedded Seyfert 1 nucleus was also inferred from the optical and X-ray observations of Isasawa et al. (1995). We note that the lack of the CO band-head could be due to a stronger continuum from the Seyfert I nucleus. XGC 7130 may be an e.xample of a collisional ring galaxy (e.g. see the review by .-Vppleton i!c Struck-Marcell 1996) formed as a result of a small companion passing through the disk of a larger system. In the radio, at 13 cm. this gala.xy is clearly extended with a linear north-south structure (L-class .source). However, the origin of the linear structure is unclear since it is coincident with a bar along the same PA. This gala.xy has been studied in detail by Shields ic Filippenko (1990), who spectrally mapped the nuclear regions of the gala.xy. Their work clearly shows the composite nature of this gala.xy. in that the line intensity ratios vary radially from those indicative of .-VGXs. close to the nucleus, to that of H II regions at increasing distances from the center. We obtained a K-band continuum infrared image of .XGC 7130 (Fig. 4.13). but no line data were obtained. In the high-surface-brightness inner regions, two tightly-wrapped spiral arms emanate from a somewhat asymmetric bar. The arms almost form a ring or lens component. The bright knot in the disk of the galaxy is a foreground star. However, the high-contrast. 148 V- VI f IS J IK IH.U -SOlfi I IS 11 trs WnwIcttKlti (Alicraii^i Wnvelrnclli (KfkrtMt^) IC SO«.l IC4WS WaTirlnielb (Mirroiul Wavelength (Mtrrwns) Figure 4.12 K-band [R spectra for [RAS 18325-5926. IC 5063. XGC 7213. and XGC 4995. Xote that \\,'e have detected H2 and Br7 in all four cases. Also detected in all cases, e.xcept IRAS 1832. is the CO 2-0 band-head absorption feature. 149 Figure 4.1;} K-banci continuum IR images of .\GC 7130. The first half of this figure is sliown as a contour plot of the intensity of the K-bancl continuum emission. It clearly shows the two tightly wrapped spiral arms that emanate from a somewhat asymmetric bar. In the second half of the figure (next page), the high-coutrast grey-scale image, one can clearly see the ring-like structure that connects with the companion to the north-west. 150 Figure 4.13 (Coatinued) 151 grey-scale image clearly shows the existence of a large faint ring of emission which is confirmed by the Digitized Sky Survey (DSS) plate of the area. small companion gala.xy lies embedded in this outer ring and the high contrast image also shows an additional faint plume to the south-west of the nuclear regions which may be connected to the outer ring. N'GC Tl.'JO has the highest FIR luminosity. log[F[R/L I ]=l 1.04. of all the gala.^ies in our sample. Much of the star-formation that powers ultra-luminous far-infrared gala.xies is believed, in part, to be triggered by interactions. While this image alone provides no evidence for the star-formation properties of this gala.xy. its morphology alone strongly suggests that it is interacting. 4.6.3 Radio Structure of IC 5063 and its Relation to Published Optical Spectra and Imaging fC 5063 is a Seyfert 2 gala.xy which is known to be abnormally radio-loud (Danziger. Goss. &; Wellington 1981: Roy .\orris 1997). We measure a flu.v density of 747 mJy at i;} cm. This gala.xy. along with IR.-VS i;J059-2407. is the furthest from and thus the most "radio-loud" with respect to. the FlR/radio correlation for our sample. .\t 3 cm the emission is in the form a linear structure composed of the bright central nucleus with two distinct knots towards the south-east, along P.A. 115±3°. Morganti, Oosterloo. Tsvetanov (1997) have also imaged this galaxy at 3 rm and our maps agree very well in morphology and flux. .-\.t 6 cm. the remnant.s of the linear structure can be seen, but a faint shell of emission is seen extending from the end of the linear structure, perhaps indicative of a bow-shock. The diffuse emission observed at (i cm may result from the interaction of the linear structure, i.e. "plasmons" ejected from the .\G.N'. with the ambient ISM in the inner disk. This gala.xy is a hybrid L+D-class (see Figure 4.3). since it is an L-class source at 3 cm and a D-class source at 6 cm. In Figure 4.14. we label the bright central source (at 3 cm) as knot .-V. the first knot to the south-east as knot B. and the second, fainter radio source as knot C. From our K-band spectra (see Fig. 4.12) we extract a log[H2/Br7] ratio of 0,14, which is confirmation of an .\G.\ in the nucleus (Mouri k Taniguchi 1992). .-VG.X-dominated FIR emission is suggested by the FIR flux ratios and the 152 q-value for this galaxy (see Table 4.5). IC 506.} is fascinating because it may be a merging system which contains one of the best- defined ionization cones known to date (Colina. Sparks, .Macchetto 1991). It is an SO gaia.xy with very peculiar dust lanes and emission-lines which define a cross-shape centered near the Colina et al. have suggested that the cross-shape has some of the characteristics of an ionization cone illuminated by an .\G.\. To e.vplain the increase in the excitation of the ionized gas with distance from the center, they suggested that this may be the result of decreasing density and a strong metallicity gradient in the central few arcsec. However, they could not completely rule out a local source of heating in the cone. Of special significance is the fact that the orientation of the linear triple is coincident with the inner edge of the ionization cone reported by Colina et al. (1991). in just the region where the ionized gas shows enhanced excitation. Knot B (see Fig. 4.14). in the 3 cm map. seems to coincide exactly with the peak in the brightness profile for both Ha and [01] A6300 emission. The fact that the peak in the emission line strength was offset from the "nucleus" was discussed at length by Colina et al. but was not explained. Flowever. these authors did note that the velocity of the peak differed by SO km s~' from the surrounding fainter gas. We will argue below that this may be a signature of ionized gas being ejected supersonically into the surrounding IS.\I in the form of plasmons. In an Hrv imaging survey of galaxies with far-infrared spectral energy distributions peaking near 60 /.im. Heisler iSc V'ader (1995) showed that this galaxy has three bright knots along, roughly, but not e.xactly, the same position angle as the radio knots seen in our -i cm image. It is interesting to note that radio knots .-V and B have similar strengths in Ha. W'e have recently obtained new IRIS .\TR long-slit spectra along the radio structure seen at cm (to be published later). The knot identified in the radio image as B contains both a .\IR continuum source and lines of [Fell] Al.644/im and Br-/. with fainter IR emitting knots on either side (this confirms the existence of the nuclear [Fell] Al.644/im detected in our spectra obtained in 1994 October). Thus, the Heisler V'ader Ha imaging and our NTR spectroscopy suggest that the nucleus is knot B. although this is by no means clear. In particular, the absolute positions lo.i i 02.0 01 ntGHT ASCENSION (J200B) KNOT A S 02.0 01 niGHT ASCENSIOfI (J200C) Figure -l.l-l -i cm radio image of IC -506.3 ovcrlayed with the cenlroid positions of the Hn knots (taken from Heisler V'ader. 1995) which are represented by crosses, (a) assumes that radio knot B is the nucleus and the Ho knot and knot B are made to align, (b) assumes that radio knot is the nucleus and the Ha knot and knot are made to align. The radio contours are 0.16 m.Jy beam"' .v (-2.5.2.5..3.4.8.10.16.20.25.30.32.40..50.64.70.90.128.256.512.800). of the optical and infrared data are uncertain by about an arcsec. and cannot therefore be accurately superimposed on the radio image. Figures 4.14a.b show that there is a slight difference in relative position between the Ha knots (shown as crosses) and the 3 cm radio knots. Figure 4.14a assumes that radio knot B is the nucleus and. hence, the Ha knot and knot B are made to align, whereas figure 4.14b assumes that radio knot is the nucleus. In each case, the other Ha knots are offset from the radio knots. We therefore speculate, regardless of whether knot .A. or knot B is the nucleus, that these offsets are consistent with plasmons being ejected into a counter-clockwise rotating disk. The plasmons ram into the ambient IS.M causing enhanced Ha emission along their leading edges. We note that if two "e.xtra-nuclear" radio blobs are indeed moving at a minimum of 80 km s~' relative to the surrounding gas. they would be highly supersonic relative to their surroundings and might naturally produce bow-shocks on their leading edge. It is therefore significant that the 6 cm map shows a bow-shock-like morphology ahead of the possible ejected plasma (such emission might be produced by electrons trapped in tangled magnetic fields in the bow-shock ahead of the ejected material). Further radio observations of higher dynamic range and spatial resolution may help to define this morphology better. Other evidence to support the view that gas outflow from the .-VG.X is interacting with the IS.M of the inner disk are broad (~600 km s~'), blueshifted absorption features seen against a disk of neutral hydrogen emission (Morganti et al. 1997). We emphasize that we are not suggesting that the overall morphology of the optical emission-line properties of IC -5063 are driven by the radio source, but rather that additional heating is occurring along one edge of the ionization cone that seems to be related to the radio structure. 4.7 Conclusions 1) Our sample contains a mi.xture of radio morphologies. There are four L-class sources, two D-class sources, one L-|-D-class. and five compact (unresolved or marginally re.solved on the arcsec scale) sources. Four of the gala.xies in the sample (.\GC 4o07. IC3639. .\GC 5I.Jo 155 and N'GC 7130) show evidence For being AGN/starburst composites, and a 5th galaxy (IC 4995) is likely to contain a compact starburst core on a scale smaller than our arc second (r< 400 pc) resolution based on its IR and radio characteristics. Further optical properties of the galaxies not studied optically here will be presented in Paper H. 2) There is a correspondence between D-class radio morphology and circumnuclear star- formation. Our optical spectroscopic observations of IC -W-id. luminous in the FIR (log[FIR/L -l] = 10.33). show clearly the importance of circumnuclear star-formation and the composite .•VG-\'/starburst nature of this galaxy. The radio emission has a clear east-west asymmetry that is seen also in the optical spectra, particularly in the (highly extended) [01] A6300 emis sion line. The radio knots appear to be contained within a region of highly extended optical [01] A630G emission. Our BPT diagrams reveal that the emission changes within the inner 1.5 kpc from .-VG.W through LI.VER. to H 11-region-like spectra radially away from the nucleus. The [01] emission that is detected to the east of the nucleus is consistent with emission that is being driven by large scale shocks (v ss 150 km s~'). perhaps being generated by the outflow of material from a nuclear starburst. In a similar vein, our optical spectroscopy of .\GC 5135 also clearly demonstrates the importance of circumnuclear star formation, .\long P.\ 19°. the emission-line ratios change radially away from the nucleus, changing from .-VCJ.N' to LI.\ER to H Il-region-like emission. 3) Some morphological segregation occurs when the galaxies global FIR colors are investi gated. For example, on a plot of IR.A.S S(100)/S(r2) versus S(60)/S(25). the C-class sources cluster to warmer IR.\S colors than the D-cIass sources. The L-class sources have a wide range of color temperatures, but the hottest gala.xies are those with well defined highly collimated radio emi.ssion. It is not clear why the radio morphology and the IR.-VS colors are related, unless the higher colors relate to the orientation of the dust torus (see for example Heisler et al (1997). and Dopita et al. (1997)). and by implication, the radio jet direction. 4) Our CTIO spectra for .N'GC 3393 and IR.A.S 11249-2859 show that the radio and op tical emission-line gas is extended on roughly the same scale. In particular we detected, in [01] Afj300. the radio knot to the south-east in IR.-VS 11249-2859. .\GC 7130. another of the 156 L-class sources, has been shown in the work of Shields Filippenko (1990) to contain a signif icant amount of circumnuclear star-formation. The L classification of this galaxy is probably due to a bar along the same PA as the radio emission. Because of the lack of optical and .\1R spectra we cannot make any conclusions about the correlation of. or lack thereof, radio mor phology with circumnuclear star-formation for the L-class sources. Future papers are planned in order to shed light on this: 5) The radio spectral indices for the core-regions of the sample are generally in a range con sistent with synchrotron emission, either from the .\G.N' itself or from cosmic rays accelerated by star-formation activity. Two notable e.xceptions are N'GC 4507 (L-class) and NGC 7213. whose low spectral indices indicate that they may suffer from free-free absorption of a thermal source. A likely source for the free-free absorption in the marginally resolved radio source .NGC 7213 might be circumnuclear star-formation, as suggested by the brightness temperature of the radio emission. Furthermore. thecoHimated structure seen in 1R.\S 11249-28.59 contains knots that are flat-spectrum sources suggesting that they might be undergoing free-free absorption; 6) IC 5003, an L-t-D class radio source, has [R.-VS fluxes and aq-value consistent with .-VGX- dominated emission, although spectroscopic confirmation of this statement is pending. The radio emission seen at 6 and 3 cm suggest that "plasmons" are being ejected from the nucleus into the disk of the gala.xy. giving rise to e.xtensive shocks, possibly a bow-shock morphology at 6 cm. and significant local heating of one edge of a well-defined ionization conc seen in optical emission lines. In the nucleus .\'IR spectra revealed Br-,. and [Fell] (along with the CO (2-0) bandhead feature): 7) Our infrared observations of IRAS 18325-5926. IC 4995. and .\GC 7213 reveal H.>. Br-;, and [Fell] AL.G44 /mi. .\lso. we detect the CO (^"=2-0) band-head absorption feature, longward of 2.3 /im. in IC 4995, IC 5063 and NGC 721.3. .-Vcknowledgments We are grateful to Russell Cannon for granting us director's time at the .A.-VT in order to obtain near-infrared spectroscopy presented in this paper. We would also like to thank the staff at both the Paul Wild Observatory and Siding Spring Observatory for their very generous 157 advice and help. Bransford thanks Sammy for his warm and friendly demeanor during visits to .\ustralia. and to R. Morganti for very valuable comments, which helped to improve the paper. We thank STSci/SERC for use of the digitized sky survey. We would also like to thank the referee for very valuable comments and suggestions that helped to improve the paper. The .Australia Telescope is funded by the Commonwealth of .Australia for operation as a .National Facility managed by CSIRO. This work is partly funded by an .\SF/L'S-.A.ustralia Cooperative Research Grant [.\T-9418142. Marston was supported by .lOVE grant .\.AG8-2G4 and a grant from .\'.\S.A. administered by the .American .\stronomical Society. Thanks to Dr. Matt Malkan for pointing out a discrepancy in the coordinates listed in Table 4.1. 4.8 References .-Vntonucci. R.. I'e Barvainis. R. 1988. .-\p.lL. Li:} .-Vppleton. P..\'.. Struck-.\Iarcell. C. 199G. in Fundamentals of Cosmic Physics. 16. Ill Baldwin. .J..A... Phillips. .\[.^[.. Terlevich. R..J. 1981. PS.\P. 9."}.5 Baum. S. O Dea. C. P.. Dallacassa. D.. de Bruyn. G.. Pedlar. .\. .-Vp.J. 199.'J. 419. 553 Bolton. .I.G.. ic Shimmins. .\..L 1973 .\u.IP.\. 30. I Colina. I... Sparks, W.B.. Macchetto. F. 1991, .Vp.I. ."570. 102 Condon. Huang. Z.P.. Yin. Q.F.. Thuan. T..\. 1991. .-Vp.!. '-iTS. (i5 Condon, .J..I.. Condon. Gisler G.. Puschell. 1982. .\p.J. 252. 102 Danziger. I..J.. Goss. W..\[.. .Jc Wellington. K..J. 1981. .\I.\R.\S. 196. 845 de Grijp. M.H.K.. Keel. W.C.. Miley. G.K.. Goudfrooij. P.. Lub. .1. 1992. .\&:.\S. 96. 389 Dopita Sutherland. R.S. 1995 .-Vp.I. 455, 468 Dopita, .\I..-\... Heisler, C.. Lumsden. S.. Bailey, .J. 1997, in preparation Evans, I.N'.. Koratkar. .\.P.. Storchi-Bergmann. T.. Kirkpatrick. H., Heckman. T.M.. Wilson. .-V.S. 1996. .\p.IS, 105, 93 Genzel. R.. Weitzcl L., Tacconi-Garman. L.E.. Blietz. .\I.. Cameron M.. Krabbe. .-\... I.utz. D.. 158 i: Sternberg, A. 1995. ApJ. 444. 129 HanifT. C.A.. Wilson. .-V-S.. Ward. M..I. 1988. .-VpJ. 334. 104 Heckman. T.M.. Gonzalez-Delgao. R.. Leitherer. C.. Meiirer. G.R.. Krolik. J.. Wilson. A.S.. Koratkar. .-V.. Kinney. A. 1997. ApJ. 482. 114 Heisler. Sc Vader. J.P. 1995. A.J. 110. 87 Heisler. C..-V.. cJc Vader. .I.P. 1994. .A..]. 107. 35 Heisler. Lumsden. S.L.. Sc Bailey. .I.A. 1997. NATURE. 385. 700 Helou. G.. Soifer. B.T.. Rowan-Robinson. M. 1985. .A.p.JL. 298. L7 [sasawa. K.. Kunieda. H.. Tawara, Y.. .\\vaki. H.. Koyamo. K.. Murayama. T.. &; Taniguchi. Y. 1995. A.I. 110. 551 Lester. D.F.. .Joy. M.. Harvey. P.M.. Ellis. H.B.. Parmer. P.S. 1987. .-Vp.J. 321. 755 Lisenfeld. U.. \'olk. H.J.. Xti. C. 1996. .\i:A. 314. 745 Morganti. R.. Oosterloo. T.. &: Tsvetanov. Z. 1997. in Emission Lines in Active Gala.xies: New Methods and Techniques. PS.\P Conference Series, vol. 113. edited by B.M. Peterson. F.-Z. Chang. A.S. Wilson, pg. 310 Mouri. H. Taniguchi. Y. 1992. .-Vp.J. 386". 68 Norris. R. P.. Allen. D. Roche, P. F. 1988. MNRAS. 234, 773 Osterbrock. D.E. 1989. in Astrophysics of Gaseous .Nebulae and .-Vctive Galactic .Nuclei. L niversity Science Books Phillips. M.M.. Charles. P..\.. ic Baldwin. -I.A. 1983. .\p.l. 266. 485 Roy. A.L.. .Norris. R.P. 1997. MN'R.-VS. in press Roy. A.L.. .Norris. R.P.. Kesteven, M..J., Troup. E.R.. I'v; Reynolds. .I.E. 1997. in preparation Rush. B.. Malkan, M..\.. &: Edelson. R..\. 1996. .\p.J. 473. 130 Savage. B.D.. Mathis. .J.S. 1979. .-\.R.\t'o.-\.. 17, 73 Shields, .J.C., k Filippenko. 1990. .\.J. 100. 1034 Stone. R.P.S.. Sc Baldwin. .J..\. 1983. M.NRAS. 204. 347 Telesco. C. M.. ic Declier. R. 1988. .\p.J. 334. 573 159 Telesco. C.M.. Bcckliri. E.E.. VVynn-Williams. C.G.. Harper. 1984. Ap.J. 282. 127 L Ivestad. .I.S.. ic Wilson. .\.S. 1989. Ap.I. 659 L'lvestacl. .I.S. 1982. Ap.J. 259. 96 Vader. .I.P.. Frogel. Terndrup. D.M.. (Sj Heisler. C.A. 199.'J. .\.J. 106. 1743 van den Broek. A.C.. van Driel. \V.. de .Jong. T.. Lub. .1.. de Grijp. M.H.K.. Goudfrooij. P. 1991. .\c^.\S. 91. 61 V'eilleiix. S.. i; Osterbrock. D.E. 1987 Ap.JS, 63, 295 Whittle. .\I.. Pedlar. .-\... Meurs. E..I..\.. L'nger. S.W.. .\.Kon. D..I.. Ward. .\I. 19-88. Ap.J. 326. 125 Wilson. .\.S.. L'lvestad. .I.S. 1982b. .\p.I. 263. 576 Wilson. A.S. 1988, Ac^A. 206. 41 Wilson. .\.S.. Ward. M..J.. HanifT. C..\. 1988. .-Vp.J. 334. 121 Wynn-Williams. C.G.. Becklin. E.E.. Scoville. .\'.Z. 1985. .A.p.1. 297. 607 160 CHAPTER 5 CONCLUSIONS 5.1 Summary of Conclusions This dissertation is focused on observational projects designed to study: i) the star forma tion of. and chemical evolution in. coHisional ring gala.xies, ii) the neutral hydrogen content (and its relation to molecular emission) and kinematics of the strongly interacting, collisional ring galaxy .\rp 118. and iii) the possible relationship between active galactic nuclei or star- burst activity and the morphology of the arc-second radio emission. In I will outline the major conclusions from the papers that constitute the core of the dissertation. In ^ I will outline ongoing research and future research plans and goals. 5.1.1 Chemical Abundances and Star Formation History of Collisional Ring Galaxies Ring gala.xics have sub-solar mctallicities in the range one-half to one-fifih solar in [0/H] and [X/H] ratios. There is a suggestion of an increase in the mean nitrogen abundance for the rings of larger linear size, but oxygen abundances show no trend with ring diameter (the oxygen abundances are approximately one-half solar). Our data does not show a decrease in the mean metallicity of the rings with increasing size, as expected from our naive assumptions. .Vaively we expected the abundances of the rings to drop off with radius, however compar ison of the o.Kvgen abundances of the ring knots with observed metallicity gradients (Zaritsky et al. 199-1: \'ila-Costas ic Edmunds 1992) reveals that this simple expectation is true only if the rings were to inhabit target galaxies of similar Hubble-type. One way of explaining the observed o.xygen abundances without rejecting completely the idea of the ring waves traveling down abundance gradients, is to suggest there is a morphological segregation of Hubble-type 161 with ring size in the progenitor galaxies (again, this suggestion arises rrom a comparison of the oxygen abundances of the rings with observed gradients). Ignoring the possibility of chemical enrichment in the ring waves, the oxygen content of the rings is consistent with them lying atop gradients in progenitors of differing Hubble-type. The smaller rings appear to inhabit small late-type galaxies (T-types 6-8) whereas the large rings appear to inhabit larger mid- to early-type galaxies (T-type 2-5). We note that this result is probably due to a selection effect. Large rings can't exist in small galaxies and small ring waves in large galaxies are most likely misidentified as tightly wrapped spiral arms or inner, hot-spot rings. The three smallest rings have the smallest .\/0 ratios whereas the larger rings have larger .\'/0 ratios. The increase of N'/O (and thus the increase of nitrogen) with ring diameter is consistent with the segregation of Hubble-type with ring size, since late-type gala.xies have lower .\/0 ratios (on average) than mid- to early-type gala.xies (V'ila-Costas Edmunds 199;}). We. however, cannot rule out the effects of heavy element production in the rings. A simple theoretical calculation suggests that there may be in situ production of oxygen in the ring waves. The observed oxygen abundances are slightly (I to 3 times) lower than predicted by our simple calculation, and may result from oxygen being produced and. i) blown out of the disk, to rain back down onto the disk some time later (i.e.. rings aren't closed boxes), or ii) hidden in a hot. unobservable 10^ Iv phase (e.g. Kobulnicky Skillman 1997). Comparison of the colors of individual starburst knots with recent numerical models of stellar evolution (Leitherer Heckman 1995: Bruzual Chariot 1995 preprint) reveal that the knots have ages that lying in the range 4 to 80 .\Iyr. These ages are less than the dynamical ages of the rings, which are typically a few x years. The fact that the knots of individual ring galaxies cluster around the same, very young age argues that their formation was recently triggered by a coherent event all around the ring. This strongly supports their production as being due to the propagation of a radial density wave. Further, the reddening corrected colors of the starburst knots in the ring gala.xies agree very well with both of the stellar evolution models. This result is very satisfying, in the sense that we have demonstrated the utility of using ring galaxies as litmus tests of theoretical models of stellar evolution. Finally. I(j2 the fact that the young star forming regions in an individual ring galaxy are coeval, and of approximately the same metallicity. makes ring systems fascinating objects for studying stellar evolution. 5.1.2 Neutral Hydrogen Observations of Arp 118 The HI emission, in addition to being highly disturbed, is distributed non-uniformly through out the disk of XGC 1144. The HI emission is highly concentrated in the XW quadrant of the ring. The CO emission is also distributed non-uniformly. and is highly concentrated in the SE quadrant of the ring. There is little or no CO emission in the nucleus, as is common in interacting gala.xies. The global HI profile has about one-third the velocity spread of the CO and Ha emission. majority of the CO emission is at high velocities (in the SE). whereas and the HI emission is at low velocities (in the WV). This is in approximate agreement with the Ho velocity field, which revealed that the H H regions were at low velocities in the .\'VV quadrant of the ring and at high velocities in the SE quadrant. Thus, the CO and HI emission appear to be "segregated", both spatially and kinematically. The molecular-to-atomic gas mass ratio. Mht/Mh[- is 0.77 in the N'W section of the ring, consistent with the observed ratios In normal spiral galaxies. This ratio is very large in the SE quadrant. 17.0. and is consistent with the observed ratios in interacting, infrared luminous galaxies. These ratios suggest that the dominant state of the interstellar medium proceeds from mainly atomic in the .\VV to molecular in the SE. In the SE quadrant, less than of the interstellar gas is in atomic form. The fact the total (HI-fH-i) hydrogen mass is approximately the same in the SE and the .\'W. but that the relative ratios of molecular-to-atomic mass change significantly, argue strongly that there is a large-scale conversion of HI-to-H* taking place in he SE region of .\GC 1144. This process probably occurs on a timescale < one rotation period (or 88 million years), which is rapid. There exists a Hl-rich dwarf gala.xy 8' to the .\E of .-Vrp 118. The emission is somewhat 163 ofT-set from the optical image of the galaxy, and extends to the S\V*. The velocity field hints at rotation, but further observations will be needed in order to better understand the kinematics of this galaxy. 5.1.3 Circumnuclear Star Formation in AGNs These galaxies contain a mixture of radio morphologies, including linear, diffuse, and com pact sources. Linear sources are thought to have their origin in a collimated outflow from an .\G.\' core, and diffuse sources have been shown to be linked with starburst activity (see Wil son 1988 .-ViiA. and references therein). Interestingly, when the FIR colors of the galaxies are considered, there is an indication that the compact radio sources have warmer FIR colors than the difi'use sources, whereas the linear sources span a wide range of FIR colors, albeit the more highly collimated sources tend to have warmer FIR colors. This morphological segregation, based on [R.\S colors, suggests that the global FIR properties may be linked in some way to the radio morphologies. Of the 5 galaxies in this sample with compact radio morphologies. 4 have radio/FIR ratios consistent with an energetically dominant AG.\. whereas one source (IC 499.5) exhibits evidence for a very compact starburst. There is a wide variation in the radio spectral-indices of the nuclei of the sample, suggesting free-free absorption is significant in a few systems, possibly arising from a vigorous starburst. The majority of the spectral indices are indicative of synchrotron emission, common in Seyfert gala.Kies. In another gala.Ky (IC 5063) we see evidence for the possible interaction of a highly collimated outflow and the surrounding rotating inner disk. Standard optical diagnostic diagrams (i.e.. Baldwin. Phillips 5: Terlevich 1981: \eilleux l'c Osterbrock 1987) revealed that the emission line ratios, for both the diffuse radio sources (IC 3639 and .\GC 5135). changes from .-VG.N'-dominated emission in the nucleus, to H Il-region- like emission in the inner 1.5 kpc. suggesting the importance of circumnuclear star formation in these systems. 164 5.2 Ongoing Research and Future Research 5.2.1 Ring Galaxies The results presented in chapter three have confirmed that the young star forming regions in an individual ring galaxy are coeval and of approximately the same metallicity. However, uncertainties in our data did not allow us to differentiate between different stellar evolution models (e.g. Leitherer Heckman 199-5: Bruzual & Chariot) which often have very important differences, i.e. the inclusion of nebular lines and models with differing metallicities. .\ddition- ally. the weak, temperature sensitive line [OHI] A4:{63 was not detected in our spectra, hence we used several empirical calibrations of the oxygen abundances with key bright lines. Therefore, we hope to obtain high resolution and signal-to-noise spectra and broad-band images to allow us to calculate ionic abundances, to confirm the trends in the abundances discovered with the current data, and to allow for a better, more precise differentiation between competing stellar evolution models. Ionic abundances can then be compared to the bright line methods, as a check of the accuracy of empirically derived abundances. Fischer et al. (1996) have clearly demonstrated the utility of far-infrared spectroscopy and near-infrared line imaging in penetrating dust enshrouded starbursts. in order to elucidate fundamental properties such as the age of the burst and the physical conditions in the ionized, warm and neutral gas. ISO observations of V'll Zw 4G6 and the Cartwheel (by \'. Charmandaris and P. .-Vppleton. private communication) have revealed patchy du.st emission. In particular, there are probably heavily obscured starburst regions in the inner ring of the Cartwheel and in a region of \'1I Zw 466 were a tidal plume is probably dumping material back onto its disk. .\GC 985 is known to contain (i.e. chapter three of this dissertation: Rodriguez Espinosa et al. 1990) areas of moderately to heavily obscured massive star formation. Far-infrared and near-infrared spectroscopy/line imaging could be used to probe these dust obscured starburst regions. Ideally, infrared line imaging could performed on a larger sample of rings, in order to: i) study their star formation history without the worry of extinction effects, ii) probe the physical conditions of large molecular clouds within the ring, iii) determine the role of shocks in 165 cloud-cloud collisions within the ring wave, and iv) map the distribution of supernova remnants. Ghigo .\ppleton (1998) have obtained high sensitivity radio continuum observations, at a number of frequencies, for a few ring systems. These observations are designed to help elucidate the nature of stellar evolution in. and behind, the ring waves. very interesting result to come of these observations is the spectral steepening of the radio continuum in .\rp 10 behind the main ring, which may relate to an aging of the relativistic electron population as a function of time. If the spectral steepening of the radio continuum is due to the aging of a population of supernova remnants, near-infrared line imaging of shock sensitive lines (particularly [Fell] Al.6-14 /mi) would be very valuable in order to map out their distribution (as has been done recently in .\[82 by Greenhouse et al. 1997) relative to the ring wave. 5.2.2 Circumnuclear Properties of AGNs In chapter two I presented the results of a pilot study designed to determine the possible relationships between .-VG.\' or starburst activity and the morphology of the arc-second radio emission. .second paper is underway to complete the optical and near-infrared spectroscopy of the sample, and to better define the star formation properties of the sample. In collaboration with P. .-Vppleton. .\I. Doptia (.\ISSSO). C. Heisler. .Marston. and R. .\orris. we have begun the COL.-V (Compact Objects in Low-power AG.\) project, which is a direct extension of the pilot study. The primary goal of the program is to determine the relationship (if any) between compact radio cores in .-VG.Ns and the circumnuclear properties of their host gala.xies. We have defined an all sky sample of 217 gala.vies. from which we will pick AG.V and non-.-\.G.\' sub-samples. The fraction of the original 217 with compact radio cores (based on V'BLI and VLB.\ snapshot observations) will be the ~.A.G.\ sub-sample.'* .\n equal number from the remaining non-radio-core gala.xies will be chosen to serve as the control sample. Hav ing carefully picked two sub-samples, detailed multiwavelength (optical, infrared, and radio) observations will be made to study their circumnuclear properties. Line imaging, long-slit spectroscopy, and broad-band imaging in the near-infrared and optical will be used to deter 166 mine liie spatial distribution of emission mechanisms including shocks, photo-ionization. and star formation, and will enable us to study internal kinematic properties of the gas. and the existence of circumnuclear structures such as rings and bars. 167 REFERENCES Appletori. P.N. 1998. iti preparation Appleton. P.N.. Sc Struck-Marcell. C. 1987a. .A.p.J. .{12. 566 Appleton. P.X.. &: Struck-Marcell. C. 1987b. Ap.I. 318. 103 Appleton. P.N.. Schombert. .J.M.. Robson. E.I. 1992. Ap.J. .385. 491 Appleton. P..\.. .Jacobs. M. 1998. in preparation .\ppleton. P.N.. Struck-Marcell. C. 1996. in Fundamentals of Cosmic Physics. 16. Ill Baldwin. .J.A.. Phillips. M.M.. Terlevich. R..J. 1981. PSAP. 93. 5 Baurn. S. O Dea. C. P.. Dallacassa. D.. de Bruyn. G.. Pedlar. .-V. .\p.J. 1993. 419. 553 Boyce. P..I.. Disney. M..I.. Boksenberg. .-V.. Crane. P.. Deharveng. .J.M.. Macchetto. F.D.. Mackay. C.D.. Sparks, W.B. 1996. 473. 760 Bruzual, Ci..-V.. Chariot. S. 1995. preprint Bushou.se. H. 19S6. .-V.J. 91. 255 Bushouse. H. 1987. Ap.I. 320. 49 Charniandaris. V'.. .\ppleton. P..\'.. Condon. .I..I.. Huang. Z.P.. Yin. Q.F.. Sc Thuan. T.X. 1991. .A.p.1. 378. 65 de \'aucouleurs. G. 1959. In Handhiich der Pliysik. vol. 53. .-Vstrophysics I\': Stellar Systems, S. Flugge (ed.). (Berlin: Spinger-Verlag). p. 275 Dopita. M..-V.. ic Evans, I.X. 1986. .A.p.1. 307. 431 Dopita M.A., i.; Sutherland. R.S. 1995 Ap.I. 455. 468 Edmunds. M.G. 19S9. in Evolutionary phenomena in galaxies. .I.E. Becknian B.E.J. Pagel (eds.). (.\e\v York: Cambridge L'niversity Press), p. 356 168 Edmunds, M.G.. Pagel. B.E.J. 1984. .\I.\R.AS. 211. 507 Fath. E..-V. 1909. Lick Observatory Bulletin. 5. 71 Fischer et al. 1996. ."Uo, L97 Garnett. D.R. 1990. -JO:}. 142 Genzel. R.. Weitzel. L.. Tacconi-Garman. L.E.. Blietz. .\I.. Cameron. M.. Krabbe. A.. Lutz. D.. Sternberg. .-V. 1995. Ap.J. 444. 129 Gerber. R..\.. Lamb. S..-\... Balsara, D.S. 1996. M.N'R.-VS. 278. ;M5 Ghigo, F.. .-Vppleton. P..\'. 1998. in preparation Greenhouse et al. 1997. .-Vp.I. 476. 105 Hecknian. T.M. 1980. .-V Heckman. T.M.. Armus. L.. Miley. G. 1990. .Ap.JS. 74. 8;W Heckman. T.M.. Gonzalez-Delgado. R.. Leitherer. C.. .Meurer. G.R.. Krolik. .J.. Wilson. Koratkar. Kinney. A. 1997. .-\.p.I. 482. 114 Hubble. E.P. 1926. .\p.I. 54. 369 Hubble. E.P. 1936. The Realm of Spiral Nebulae. (Xew Haven: Yale University Press) Kennicutt. R.C.. < Kennicutt. R.C.. Roettiger. K..-V.. Keel, VV'.C.. Van Der Hulst. <.V; Hummel. E. 1987. .-V.I. 93. mil Kennicutt, R.C.. Keel. VV'.C.. Blaha. C..-V 1989. .-V.J. 97. 1022 Kobulnicky. H..-V.. Skillman E.D. 1997. .-VpJ. 489. 636 Larson. R.B.. Tinsley B.M. 1978, -Vp.I. 219 46 Leitherer. C.. 1*1: Heckman. T..VL 1995, .VpJS, 96, 9 Lynds, R., Toomre. .-V. 1976 .-Vp.J. 209. .382 .Marston. .-V.P., ic .-Vppleton. P.X. 1995. -V.I. 105. 1002 .\IcCall. .M.L.. Rybski. P..\I.. ic Shields. G..-V. 1985, .-Vp.JS. 57. 1 -Miles, .I.VV'.. Houck. .I.R.. Hayward. T.L.. .-Vshby. .\LL..X. 1996. .-Vp.J, 465, 191 169 Osterbrock. D.E. 1989. Astrophysics of Gaseous Sebulae and Active Galactic S'uclei (Mill Valley. CA: L'niversity Science Books) Pagel. B.EJ.. Edtnunds. M.G.. Blackwell. D.E.. Chun. M.S.. i.: Smith. G. 1979. MN'RAS. 189. 95 Penston. M.V.. Perez. E. 1984. MNRAS. 211. ;5:} Peterson. B.M.. Wanders. [. 1998. to be published in Structure and Kinematics of Quasar Broad Line Regions (San Francisco; .-Vstronomical Society of the Pacific) Peterson. B.M. 1997. An Introduction to Active Galactic Nuclei (New York. N'^'r Cambridge L'niversity Press) Pogge. R.W. 19S8. ApJ. 328. 519 Rodriguez Espinosa. J.M.. L Stanga. R.M. 1990, ApJ. 365. 502 Sandage. A. 1961. The Hubble Atlas of Galaxies (Washington: Carnegie Institution) Seyfert. C.K. 1943. ApJ. 97. 28 Shields. J.C.. Filippenko. .\.V. 1990. AJ. 100. 1034 Skillman. E.D. 1998. to be published in \'lll Canary Islands Winter School of Astrophysics: Stellar Astrophysics for the Local Groups. A. Aparico. .\. Herrero. io F. Sanchez (eds.). (Xevv York: Cambridge L'niversity Press) Slipher. \'.M. 1917. Lowell Observatory Bulletin. 3. 59 Soifer. B.T.. Sanders. D.B.. Madore. B.F.. Xeugebauer. G.E.. Danielson. G.E.. Elias. J.H.. Lonsdale. C.J.. Rice. W.L. 1987. .\pj. 320. 238 Struck-.\Iarcell. C.. Struve Sc Zerbergs 1962. .Astronomy of the 20"^ Century (.\'e\v York. N Y: The Macmillan Company) Theys. J.C.. Spiegel. E..-\. 1976. .\pj. 208. 650 Theys. J.C.. io Spiegel. E.A. 1976. ApJ. 212. 616 Toomre. 1978. in The Structure and Evolution of Galaxies. l.\L' Symp. No. 79. M.S. Longhair and J. Einasto (eds.). (Holland: D. Reidel Publications), p. 109 Toomre, A., Toomre. J. 1972. .ApJ. 178. 623 \'eilleux. S.. Osterbrock. D.E. 1987 .A.pJS. 63. 295 170 V'ila-Costas. M.B.. &: Eclnuincis. M.G. 1992. MN'R.-VS. 2.59. 121 \ ila-Costas. M.B., Eclmimds. .\I.G. 199;}. .VINRAS, 265. 199 Vilchez. .I.-VI. 1988. Ph.D. Thesis. La Laguna. Tenerife Wilson. .\.S. 1988. .-VijA. 206. 41 Zaritsky. D.. Kennicutt. R.. Huchra, .I.P. 199-1. .Ap.!. -120. 87 IMAGE EVALUATION TEST TARGET (QA-3) 150mm IIVUGE.Inc 1653 East Main Street Rochester. NY 14609 USA Phone: 716/482-0300 Fax: 716/288-5989 O 1993. ApfXted Image. Inc. All Rights Reserved II Zw 28 (D CO in a» knot 1 o 10 & ^ oCO w + I i Ll 0 a J I I I I I 1 L J L_J I L_J I I I l__J I I I J I I I I I I L_J I I o 3500 4000 4500 5000 5500 6000 6500 7000 (0 0)Ih