Multi-dimensional analysis of the chemical and physical properties of spiral galaxies
aa r X i v : . [ a s t r o - ph . C O ] J un M ULTI - DI M EN S IO NA L ANALYS IS OF THECHEMICAL AND P HYS ICAL P ROP ERTIESOF S P IRAL GALAXIES
Thesis submitted for the degree ofDoctor of PhilosophybyF
ERNANDO F ABI ´ AN R OSALES O RTEGA
Institute of Astronomy T RINITY C OLLEGE
University of Cambridge
November 5, 2009
UMMARY
The emergence of a new generation of instrumentation in astrophysics, which provide spatially-resolvedspectra over a large 2-dimensional (2D) field of view, offers the opportunity to perform emission-line surveysbased on samples of hundreds of spectra in a 2D context, enabling us to test, confirm, and extend the previousbody of results from small-sample studies based on typical long-slit spectroscopy, while at the same timeopening up a new frontier of studying the 2D structure of physical and chemical properties of the disks ofnearby spiral galaxies. The project developed in this dissertation represents the first endeavour to obtainfull 2D coverage of the disks of a sample of spiral galaxies in the nearby universe, by the application ofthe Integral Field Spectroscopy (IFS) technique. The semi-continuous coverage spectra provided by thisspectral imaging technique allows to study the small and intermediate linear scale variation in line emissionand the gas chemistry in the whole surface of a spiral galaxy.The PPAK IFS Nearby Galaxies Survey: PINGS, was a carefully devised observational project, designedto construct 2D spectroscopic mosaics of 17 nearby galaxies in the optical wavelength range. The sampleincludes different galaxy types, including normal, lopsided, interacting and barred spirals with a good rangeof galactic properties and star forming environments, with multi-wavelength public data. The spectroscopicdata set comprises more than 50 000 individual spectra, covering an observed area of nearly 100 arcmin ,an observed surface without precedents by an IFS study. All sources of errors and uncertainties during thereduction process of the IFS observations are assessed very carefully. This methodology contributed not onlyto improve the standard reduction pipeline procedure for the particularly used instrument, improvements thatcan be applied to any similar integral-field observation and/or data reduction, but to defining a self-consistentmethodology in terms of observation, data reduction and analysis for the kind of IFS surveys presented inthis dissertation, as well as providing a whole new set of IFS visualization and analysis software madeavailable for the public domain.The scientific analysis of this dissertation comprises the study of the integrated properties of the ionizedgas of the whole PINGS sample, and a detailed 2D study of the physical and chemical abundance distribu-tion derived from the emission line spectra of four selected galaxies of the sample. Spatially-resolved mapsof the emission line intensities and physical properties are derived for each the selected galaxies. Differentmethodologies are explored in order to study the spatially-resolved spectroscopic properties of the galaxies.Abundance analysis are performed based on a variety of diagnostic techniques using reddening correctedspectra. From this analysis, evidence is found to support that the measurements of emission lines of a “clas-sical” H II region are not only aperture, but spatial dependent, and therefore, the derived physical parametersand metallicity content may significantly depend on the morphology of the region, on the slit/fibre position,on the extraction aperture and on the signal-to-noise of the observed spectrum. On the other hand, the resultspresented in this dissertation indicate the existence of non-linear multi-modal abundance gradients in normalspiral galaxies, consistent with a flattening in the innermost and outermost parts of the galactic discs, withimportant implications in terms of the chemical evolution of galaxies.The powerful capabilities of wide-field 2D spectroscopic studies are proven. The chemical compositionof the whole surface of a spiral galaxy is characterised for the first time as a function not only of radius, but ofthe intrinsic morphology of the galaxy, allowing a more realistic determination of their physical properties.The methodology, analysis and results of this dissertation will hopefully contribute in a significant way tounderstand the nature of the physical and chemical properties of the gas phase in spiral galaxies. iii ECLARATIO N
I hereby declare that this thesis entitled
Multi-dimensional analysis of the chemical and physical propertiesof spiral galaxies is not substantially the same as any that I have submitted for a degree or diploma or otherqualification at any other University. I further state that no part of this dissertation has already been or isbeing concurrently submitted for any such degree, diploma or other qualification. This thesis is essentiallythe result of my own work, and includes nothing which is the outcome of work done in collaboration exceptwhere specifically indicated. Those parts of this dissertation which are undergoing review for publication,or included in conference proceedings are as follows: • Chapter 3
The work presented in this chapter is essentially my own, it has been submitted for publication as:Rosales-Ortega, F. F., Kennicutt, R. C., S´anchez, S. F., D´ıaz, A. I., Pasquali, A., Johnson, B. D. andHao, C. N., (2009)
PINGS: the PPAK IFS Nearby Galaxies Survey , submitted to Monthly Notices ofthe Royal Astronomical Society, and it was benefited from collaboration with these authors. • Chapter 4
The work presented in this chapter is essentially my own, a large part of it was submitted for publi-cation in the article mentioned above. The work related to NGC 628 has been partially done withina collaboration, and submitted for publication as: S´anchez, S. F., Rosales-Ortega, F. F., Kennicutt,R. C., D´ıaz, A. I., Pasquali, A., Johnson, B. D. and Hao, C. N., (2009)
PPAK Wide-field Integral FieldSpectroscopy of NGC 628: The largest spectroscopic survey on a single galaxy , submitted to MonthlyNotices of the Royal Astronomical Society. The
Gaussian-suppression and absolute fux calibrationtechniques were developed in collaboration with S. F. S´anchez. Their implementation and initial test-ing were carried out in parallel by both of us, though the final implementation for this dissertation aremy own work. • Chapter 5
The work presented in this chapter is essentially my own. Some parts has been partially done withina collaboration, and submitted for publication in the articles mentioned above. • Chapter 6
The work presented in this chapter is essentially my own, but benefited from discussions and advicefrom S. F. S´anchez and A. I. D´ıaz. • Some figures of chapter 3 and chapter 6 have been included in
The Promise of Multiwavelengthand IFU observations , Kennicutt, R. C., Hao, C. N., Johnson, B. D., Rosales-Ortega, F. F., D´ıaz,A. I., Pasquali, A. and S´anchez, S. F., (2009) Proceedings of the IAU Symposium 262, G. Bruzual,S. Charlot, eds.This thesis is less than 60 000 words in length.Fernando Fabi´an Rosales Ortega
Cambridge, November 5, 2009 v HES IS CONTENT
The whole thesis is not included in astro-ph due to file size limitations.The full contents can be found at: vii
Introduction T he existence and distribution of the chemical elements and their isotopes in the universe is aconsequence of very complex processes that have taken place in the past since the Big Bang andsubsequently in stars and in the interstellar medium (ISM) of the present day galaxies, wherethey are still ongoing. These processes have been studied theoretically, experimentally and ob-servationally. Different theories of cosmology, stellar evolution and interstellar processes have been consid-ered, laboratory investigations of nuclear and particle physics, studies of elemental and isotopic abundancesin the Earth and meteorites have also been involved, as well as astronomical observations of the physicalnature and chemical composition of stars, galaxies and the interstellar medium.From the observational point of view, the study of chemical abundances in galaxies, like many otherareas of astrophysics, has undergone a remarkable acceleration in the flow of data over the last few years.We have witnessed wholesale abundances determinations in tens of thousands of galaxies from large scalesurveys such as the Two Degree Field Galaxy Redshift Survey (2dFGRS, Colless et al., 2001) and the SloanDigital Sky Survey (SDSS, York et al., 2000), measurements of abundances in individual stars of LocalGroup galaxies beyond the immediate vicinity of the Milky Way, and the determination of the chemicalcomposition of some of the first stars to form in the Galactic halo. Chemical abundances studies are alsoincreasingly being extended to high redshift, charting the progress of stellar nucleosynthesis over most ofthe age of the universe. The primary motivation common to all of these observational efforts is to use thechemical information as one of the means at our disposal to link the properties of high redshift galaxies withthose we see around us today, and thereby understand the physical processes at play in the formation andevolution of galaxies.The galactic chemical evolution is dictated by a complex array of parameters, including the local initialcomposition, star formation history (SFH), gas infall and outflows, radial transport and mixing of gas withindisks, stellar yields, and the initial mass function (IMF). Although is difficult to disentangle the effects ofthe various contributors, measurements of current elemental abundances constrain the possible evolutionaryhistories of the existing stars and galaxies. Important constraints on theories of galactic chemical evolution1 Chapter 1. Introductionand on the star formation histories of galaxies can be derived from the accurate determination of chemicalabundances either in individual star-forming regions distributed across galaxies or through the comparison ofabundances between galaxies. Nebular emission lines from individual H II regions have been, historically,the main tool at our disposal for the direct measurement of the gas-phase abundance at discrete spatialpositions in low redshift galaxies.However, in order to obtain a deeper insight of the mechanisms that rule the chemical evolution of galax-ies, we require a significantly the number of H II regions sampled in any given galaxy. In this dissertation,I present a new observational technique conceived to tackle the problem of the 2-dimensional coverage ofthe whole surface of a galaxy. The advent of new spectroscopic techniques provides powerful tools forstudying the small and intermediate scale-size variation in line emission and stellar continuum in nearbywell-resolved galaxies. In this work, I address the problems and challenges that imply the determinationof the chemical composition in galaxies in a 2D context and the subsequent derivation of their physicalproperties.I will begin by presenting in this chapter a literature review on the determination of chemical abundancesin galaxies. As an introduction to this topic, the physics of gaseous nebulae is discussed in § II regions in § § § II extragalactic regions by Dinerstein (1990), P´erez-Montero& D´ıaz (2005), Tielens (2005), and Osterbrock & Ferland (2006). This discussion leads to the presenta-tion of new techniques and methods for the determination of chemical abundances in nearby galaxies asdescribed in § Gaseous Nebulae are observed as bright extended objects in the sky, some are easily observed on directimages but many others are intrinsically less luminous or are affected by interstellar extinction on ordinaryimages, but can be resolved on long exposures with special filters and techniques so that the background andforeground stellar and sky radiations are suppressed. The surface brightness of a nebula is independent of itsdistance, but more distant nebulae have (on average) smaller angular size and greater interstellar extinction.Gaseous nebula have emission-line spectra. Nebulae emit electromagnetic radiation over a broad spectralrange, although only a few wavelengths pass easily through the Earth’s atmosphere. Visible light and someinfrared and radio radiation can be studied from the ground, but most other wavelengths can only be coveredfrom high-latitude aircrafts or space telescopes. The source of energy that enables normal emission nebulaeto radiate is ultraviolet radiation from stars within or near the nebula. There should be one or more starswith effective surface temperature T ⋆ ≥ K, the ultraviolet photons of these stars transfer energyto the nebula by photoionization. In all nebulae, hydrogen (H) is by far the most abundant element, andphotoionization of H is thus the main energy-input mechanism. Photons with energy greater than 13.6eV (the ionization potential of H), are absorbed in the process, and the excess energy of each absorbedphoton over the ionization potential appears as kinetic energy of a newly liberated photoelectron. Collisionsbetween electrons and between electrons and ions, distribute this energy and maintain a Maxwellian velocitydistribution with temperature T in the range 5,000 < T < n e ≤ cm − ) oftypical nebulae, collisional de-excitation is even less probable, so almost every excitation leads to emissionof a photon, and the nebula thus emits a forbidden-line spectrum that is quite difficult to reproduce underterrestrial laboratory conditions.Thermal electrons are recaptured by the ions, and the degree of ionization at each point in the nebula isfixed by the equilibrium between photoionization and recapture. In the recombination process, recapturesoccur to excited levels, and the excited atoms thus formed then decay to lower and lower levels by radiativetransitions, eventually ending in the ground level. In this process, line photons are emitted and this is theorigin of the H I Balmer and Paschen line spectra observed in all gaseous nebulae. The recombinationof H + gives rise to excited atoms of H and thus leads to the emission of the H I spectrum. Likewise, He + recombines and emits the He I spectrum, and in the most highly ionized regions, He ++ recombines and emitsthe He II spectrum. Recombination lines of trace elements are also emitted; however, the main excitationprocess responsible for the observed strengths of such lines with the same spin or multiplicity as the groundterm is resonance fluorescence by photons, which is much less effective for H and He lines because theresonance lines of these more abundant elements have greater optical depth. Nevertheless, line emission ofthese rare elements plays a significant role in the physics of the nebula, and permits the determination of thechemical composition inside the nebula.The spectra of gaseous nebulae are dominated by collisionally excited forbidden lines of ions of com-mon elements, such as [O III ] ll II ] ll II ] ll II ] ll l al b l g l I l II l II , C III , C IV and so on. Nebular emission-line spectra extend into other spectral ranges, in the infrared forexample, the [Ne II ] l m m and [O III ] l m m are among the strongest lines measured, into the ultra-violet Mg II ll III ] ll IV ll a l I at l > < l < II regions and planetary nebulae (PNe). Thoughthe physical processes in both types are quite similar, the two groups differ greatly in origin, mass, evolutionand age. The objects of study for the chemical composition of galaxies are extragalactic H II regions. Thesediffuse nebulae are regions of interstellar gas in which the exciting stars are O- or early B-type stars, i.e.young stars which use up their nuclear energy quickly. These hot, luminous stars undoubtedly formed fairlyrecent from interstellar matter that would be otherwise be part of the same nebula. The effective temperature Chapter 1. Introductionof the stars are in the range 3 x 10 < T ⋆ < K; throughout the nebula, H is ionized, He is singlyionized and other elements are mostly singly or doubly ionized. Typical densities in the ionized part ofthe nebula are of the order 10 or 10 cm − , ranging to high as 10 cm − . In many nebulae, dense neutralcondensations are scattered through the ionized volume. Internal motions occur in the gas with velocitiesof order 10 km s − , approximately the isothermal sound speed. Bright rims, knots, condensations, and soon, are apparent to the limit of resolution. The hot, ionized gas tends to expand into the cooler surroundingneutral gas, thus decreasing the density within the nebula and increasing the ionized volume. The outer edgeof the nebula is surrounded by ionization fronts running out into the neutral gas. This original two-phasemodel of the interstellar medium (the H II /H I region dichotomy) was introduced by Str¨omgren (1939). Heshowed that photoionized gas near hot stars is segregated into physically distinct volumes, separated fromtheir neutral environment by sharp boundaries. II regions The spectra of H II regions are strong in H I recombination lines, [N II ] , [O II ] and [O III ] collisionally excitedlines, but the strengths of N and O may differ greatly, being stronger in the nebulae with higher central-star temperatures. The brightest H II regions can easily be seen on almost any large-scale image of nearbygalaxies, and those taken in a narrow wavelength band in the red (including H a and [N II ] lines) are speciallyeffective in showing faint and often heavily obscured extragalactic H II regions. The H II regions are stronglyconcentrated to the spiral arms and indeed are the best objects for tracing the chemical composition, structureand dynamics of the spiral arms in distant galaxies. They trace recent star formation and, through the analysisof their chemical composition, previous star formation activity. Typical masses of observed H II regions areof the order 10 to 10 M ⊙ , with the lower limit depending largely on the sensitivity of the observationalmethod used.H II regions are the only form of interstellar material which emits strongly in the optical spectral region;therefore, there is a much longer and richer history of observations and theory for them than for the otherthermal phases of interstellar matter. Optical observations of H II regions provide fairly complete informa-tion about their elemental composition. From their spectra, abundances relative to hydrogen can be estimatedfor nearly all of the most common elements, particularly He, N, O, Ne, Ar, and S (note that oxygen aloneconstitutes nearly 50% by mass of the elements heavier than helium.) Furthermore, ionized nebulae areremarkably efficient machines for converting ultraviolet continuum energy from OB stars, originally dilutedover wide bandpasses, into a few narrow, intense, optically-thin emission lines. The intrinsic emissivitiesof these lines are easy to calculate in principle, although they are sensitive to the local thermodynamic stateof the gas (electron density n e and temperature T e ). On the other hand, the thermal parameters can also bedetermined from the spectra, using diagnostic line-intensity ratios. In this way, H II regions can be used tomeasure element abundances in the (present-day) gas of distant galaxies.The sample of extragalactic H II regions studied so far has metal abundances ranging from about 0.02 toseveral times solar. This is a useful complement to studies of our own Galaxy, which contains no severelymetal-deficient H II regions (except for a handful of planetary nebulae formed by stars of the halo popu-lation). In contrast, for many H II regions in the outskirts of late-type spirals and in some dwarf irregulargalaxies, the process of metal enrichment by stellar nucleosynthesis is still in its early stages, providing ahint on the early chemical evolution of galaxies. These low-metallicity H II regions are also presumed tohave experienced only a small degree of alteration in their helium abundances due to stellar activity. There-.1ThePhysics of Gaseous Nebulae 5fore, their present He/H ratios should be nearly the same as the primordial value, providing valuable testsfor cosmological theories. The various categories of extragalactic H II regions are essentially lists of theirenvironments. These include:1. Disk H II regions in spiral and irregular galaxies.2. Gassy dwarf irregular galaxies with spectra which are heavily dominated by H II regions.3. Nuclear and near-nuclear regions sometimes called “starburst” or “hotspot” H II regions (e.g. Kenni-cutt et al., 1989).The first two categories have the best abundance data available in the literature. Regions in the thirdgroup tend to have relatively strong stellar continua and to be fairly metal-rich, which make it to difficultto obtain accurate measurements of the emission lines from which abundances are determined. On theother hand, members of the first two categories are universally regarded as members of the same family.H II regions in nearby galaxies have been well-catalogued; atlases are available for the Large and SmallMagellanic Clouds (LMC, SMC), and a large number of other galaxies (e.g. Hodge & Wright, 1967, 1977;Hodge & Kennicutt, 1983). The star-forming dwarf irregulars are usually found by spectroscopic surveysfor emission-line galaxies (e.g. Kinman 1984 and more recently the SDSS data releases).The statistical properties of the H II region populations in spiral and irregular galaxies were addressed byKennicutt (1988) and Kennicutt et al. (1989). They find that late-type galaxies have both intrinsically higher-luminosity H II regions, and larger total numbers of H II regions after normalization by galaxy size, than doearly-type spirals. Within a galaxy, the differential luminosity function of the H II regions is roughly apower-law, N (cid:181) L − ± . , although some low-luminosity irregulars have an exceptional supergiant complex,and Sa-Sb galaxies are deficient in luminous regions. While the positive correlation between the luminosityof the brightest H II region and that of the parent galaxy can be understood as chiefly a sample-size effect,the dependence on morphological type is a real and separate factor. Typical large galaxies contain hundredsof optically detectable H II regions. It is important to note that of all the regions detected and cataloged inH a or H b , it is usually the nearest and the most luminous “giant” H II regions for which abundances arederived.Some of the best-studied regions are the 30 Doradus complex in the LMC, NGC 604 in M 33, andNGC 5461 and 5471 in M 101. Selection effects play an important role, necessarily poorer spatial resolutioncontributes to a tendency to identify larger regions in more distant galaxies. This effect is illustrated by Israelet al. (1975), who compare large-beam radio measurements with optical images of the same H II regions inM 101; at better resolution these regions break up into groups or chains of smaller clumps. Likewise, H II re-gions in dwarf irregulars are also found to have complex structure when closely examined (e.g. Hodge et al.,1989; Davidson et al., 1989). In more distant galaxies, we will always be looking at more heterogeneousvolumes; for example, a typical aperture size (4”) for spectrophotometric studies corresponds to 1 pc at 50kpc (the LMC) and 2 kpc at 100 Mpc.The morphology of many giant extragalactic H II regions can be characterized to first order as a “core-halo” structure, on the basis of both optical and radio-continuum data. The cores are composed of densematerial, often in several distinct clumps, close to the ionizing stars. The diffuse, lower-density envelopes arepresumably ionized by photons escaping from the inner regions and represent the radiation-bounded edgesof the Str¨omgren volume. Most giant extragalactic H II regions are believed to be essentially radiation-bounded (e.g. McCall et al., 1985). In addition, the denser regions themselves are inhomogeneous, as seen Chapter 1. Introductionin the detailed studies of NGC 5471 by Skillman (1985), and of NGC 604 by Diaz et al. (1987). That thereare also inhomogeneities on smaller spatial scales is shown by the discrepancy between (rms) n e valuesderived from recombination emission and local values determined from density-sensitive line ratios. Thedense clumps are embedded in a much lower-density medium, with typical clump volume filling factors of0.01 – 0.1 (e.g. Kennicutt, 1984; McCall et al., 1985). The interclump material is often treated as a vacuumin nebular models, because it does not contribute significantly to the optical emission lines.Giant extragalactic H II regions display supersonic velocities, which appear to correlate with H b lumi-nosity. Terlevich & Melnick (1981) interpret the line-widths as virial and therefore usable for determiningthe local gravitational field; they also find a secondary dependence on metallicity. An alternative interpre-tation of the origin of the line-widths is that they are a result of stellar winds from the exciting stars, andpossibly also from embedded supernova remnants (e.g Dopita, 1981; Skillman, 1985). For nearby regions,it is possible to actually identify the stars which may be responsible for driving the high-velocity gas.As mentioned above, luminous extragalactic H II regions are ionized by OB associations. For nearbyregions, the members of the stellar cluster can be distinguished individually and HR diagrams can be con-structed. The nebular ionization structure and emitted spectrum will evolve as the cluster ages and the UVradiation field diminishes and softens. Wolf-Rayet stars are often present in extragalactic H II regions, thefrequency of Wolf-Rayet stars is higher for higher-metallicity regions as proved by Maeder et al. (1980).Wolf-Rayet stars are important in this context because they furnish metal-rich outflows which are capableof altering the chemical composition of their gaseous environment. Giant H II regions are also known hostsof Type II supernovae. Winds and supernovae from the massive stars can contaminate (or enrich) the localgas in H II regions in He, C, O, and other species. Evidence for such local enrichments has been sought andperhaps seen in some regions (Skillman, 1985; Pagel, 1986). II regions H II regions are ideal places to determine the abundance of the elements that are responsible for recombina-tion and fine-structure lines. The list of these elements is generally limited at the present time, although linesof many more elements are observable in several planetary nebulae and high spectral resolution and multi-wavelength studies of nearby H II regions (e.g. Garc´ıa-Rojas et al., 2006). The determination of elementabundances in H II regions are given relative to the hydrogen content, which is observed by its recombina-tion lines. Only a few abundant elements give observable recombination lines with similar physics: helium,carbon, nitrogen and oxygen, whose lines are very weak to detect and may suffer problems with fluores-cence excitation. The abundances derived from the fine-structure lines in the visible are sensitive to bothtemperature and density, and the interpretation of the line intensities is a delicate problem. In some cases,the temperature of the emitting zone can be obtained and the abundance determination is safer. However,even in this case temperature fluctuations can yield systematic errors in the abundances.The abundances derived from the mid-and far-infrared fine structure lines are not sensitive to electrontemperature and are little affected by extinction. These are considerable advantages with respect to theoptical lines. However, there are important discrepancies between the abundances derived from infraredand from optical lines. These differences may originate in temperature fluctuations or in errors in atomicparameters, but one has to consider that the critical density for infrared lines is generally much smaller thanfor visible lines, so that the abundances derived from the infrared lines are underestimated if the density ishigh (this effect can be important for planetary nebulae)..2Determination ofchemical abundances in H II regions 7 Observed emission line intensitiesReddening-corrected line intensitiesElement abundance ratios (O/H, etc.)Local physical conditions (n , T ) e e
Ionic abundance ratios (O /H , O /H , etc.) + + ++ +
Correction for extinction using hydrogen recombination decrementDiagnostic line ratios, e.g. [O III] for T , [S II] for n e e
Ionic level populations and calculated line emissivitiesCorrection for unobserved ions
Figure 1.1:
The direct method of chemical abundance determinations.
All elements (with the obvious exception of H) exist in several ionization states in H II regions. However,only the abundances of those ions that emit observable lines can be determined. If such ions are minorspecies, they yield no useful information because the physical parameters of H II regions are most oftentoo uncertain to allow an accurate solution of the ionization equilibrium. This is for example the case forO I , C II , S II or Si II . If the observed ion is a major species the situation is more favorable since we cancalculate, more or less accurately, the abundances of the unobserved ions of the same element. However,uncertainties remain if a high precision is required, as is the case of helium in a cosmological context. Themost favorable case is that of oxygen, whose major ionization states, O II and O III , are observable opticallyand for which the electron temperature T e can be determined. For this reason, oxygen is, after helium, theelement whose abundance is best determined, at least if the temperature is large enough (or the metallicityis too low) for the temperature-sensitive lines (e.g. [O III ] l II region to estimate theelectron temperatures and ionization correction factors for individual ions.However, given the difficulty of detecting the T e -sensitive line and the assumptions made in nebular mod-eling, a very popular approach is to obtain the abundance of extragalactic H II regions using empirical rela-tions between the oxygen abundance and the intensity of the [O II ] ll III ] ll b (Pagel, 1997) or by using the [O II ] ll II galaxies. This method however, is the less accurate andmuch discussion about the reliability of the different empirical calibrations is still ongoing in the literature(e.g. see Kewley & Ellison, 2008, for a thorough discussion). A full discussion regarding this topic is beyondthe scope of this chapter, however, in ?? , I include a small review on the different empirical techniques ofabundance determinations (considering their particular advantages and pitfalls), and their implementationsin the context of the work carried out in this dissertation. A more complete explanation of the determinationof nebular abundances from emission lines can be found in references on the physics of gaseous nebulaesuch as Aller (1984), and Osterbrock & Ferland (2006). It has been noticed that certain H II region emission-line ratios, such as [O III ]/H b , vary across the disks ofnearby spiral galaxies. The interpretation of this variation in terms of a metallicity trend was introducedby Searle (1971), in a paper that laid the groundwork for the entire field of abundance gradients. It wassoon followed up by further observational studies and a more rigorous analysis involving the constructionof realistic nebular models (Shields, 1974). From the start it was recognized that there was a need for a“second parameter” in addition to the O/H ratio, to explain an observed systematic increase in O ++ /O + withdecreasing O/H. Shields & Tinsley (1976) suggested that this secondary effect results from a tendency forthe effective temperatures of the ionizing stars to be hotter for lower O/H, and interpreted it as a metallicity-dependent truncation of the top end of the initial mass function (i.e. that the formation of very massive stars isinhibited by higher metallicity). Some form of the idea of a Z -dependent IMF is still a popular interpretationof the “excitation” trend (e.g. Vilchez & Pagel, 1988), but it is also the case that a similar effect can arisefrom systematic variations in the nebular geometry and/or filling factor (Mathis, 1985; Dopita & Evans,1986).An extensive body of literature has been amassed on the subject of abundance gradients in galaxies. Notsurprisingly, many works have focused on large, nearby galaxies with many observable H II regions, suchas M 33 (Vilchez et al., 1988; Rosolowsky & Simon, 2008) and M 101 (Evans, 1986; Torres-Peimbert et al.,1989; Kennicutt & Garnett, 1996). The gradients are usually expressed as a logarithmic fit to some 5 – 20regions per galaxy, and have a magnitude of about d log(O/H)/R = -0.08 ( ± II region samplesare often small, and more fundamentally because these are generally the most metal-rich regions, for which[O III ] l II regions are intrinsically fainter, but studies of M 81 (Sab) show it to have an O/H gradientsimilar to those of M 33 and M 101 (Garnett & Shields, 1987). There is at present no convincing evidencethat the O/H gradient depends on morphological type among spiral galaxies. However, there is evidence fora good correlation between mean O/H abundance and the overall galaxy mass or luminosity. This trend re-sembles the correlation of stellar metallicity with galaxy mass, and probably has its roots in the fundamentalprocesses of galaxy formation and evolution.Along with the trend in [O III ]/H b , a similar radial trend was noted for the ratio [N II ]/H a , which de-.3Abundance gradients ingalactic disks 9 Figure 1.2:
The O/H and N/H abundance vs. galactocentric distance in M 33, examples of the radial oxygenand nitrogen abundance gradients. Plots taken from Magrini et al. (2007). creases with increasing distance from the centers of spiral galaxies. Although part of this trend is due to thegenerally lower degree of ionization in the outer H II regions, there also must be a real variation in abun-dance. Unlike oxygen, for nitrogen one usually can measure the singly-ionized state only; unfortunately,N ++ has no strong optical lines. As mentioned before, the nitrogen abundance is basically derived from[N II ]/[O II ]. The relative behavior of O and N is often displayed by plotting N/O vs. O/H. Some studiesfind that N/O varies almost as steeply as O/H, which has special significance in the context of chemicalevolution models, but others claim that N/O varies only slightly or is constant across the disks of galaxiessuch as M 101, M 33, M 81, and M 83. There also appear to be variations in N/O at a given O/H fromgalaxy to galaxy (same references as above). Some of these variations may be an artifact of the analysis,especially since N + contains only a small fraction of the nitrogen for the lowest-abundance, most highly ion-ized regions. For such regions, the ionization correction factors are very large, and the uncertainties in theionization structure translate into large uncertainties in the elemental abundance of nitrogen. Nevertheless,there is accumulating evidence that nitrogen has a more complicated behavior than does oxygen, with N/Obeing roughly constant at low values of O/H and increasing at higher O/H (e.g. Pagel, 1985; Torres-Peimbertet al., 1989). Measurements of N/O in metal-poor dwarf irregular galaxies are an important ingredient inthis argument.Scatter in gradient determinations has been seen in various studies (e.g., in the Milky Way Afflerbachet al. 1997 or in M 33 Rosolowsky & Simon 2008), even after accounting for uncertainties in the stellarabsorption and reddening corrections, an intrinsic scatter of ∼ Because of interstellar extinction, one can use the same techniques as for extragalactic H II regions onlyfor the part of our Galaxy outside a galactocentric distance of about 7 kpc. Studies such as those by Haw-ley (1978) found gradients similar to those in other spirals, d log(O/H)/ d R = -0.04 to -0.06 dex/kpc and d log(N/H)/ d R = -0.10 dex/kpc. Determination of abundances in the inner galaxy requires the use of othertechniques, such as measuring electron temperatures from radio recombination lines. The values of T e arefound to increase systematically with increasing radius, presumably because of a decreasing abundance ofoxygen, the primary coolant. The inferred gradient in O/H is d log(O/H)/ d R = -0.07 dex/kpc after the classicpaper of Shaver et al. (1983).The results from optical studies for the other measurable elements are similar to those for other galaxies:N/H varies more steeply than O/H; S/O, Ne/O, and Ar/O do not vary in the outer part of the Galactic disk.Again, the optical studies are restricted to the unobscured portion of the Milky Way galaxy, and thereforedo not sample the inner disk where the inferred O/H values are high. A more recent development, madepossible by improvements in infrared detectors and the availability of space observatories. The explorationof the infrared spectral region as a tool for studying the galactic abundance gradient. The mid-infraredspectral region (5-30 m m) contains emission lines of the major ions of Ar, S, and Ne: [Ar II ] 7.0 and [Ar III ]9.0 m m; [S III ] 18 and [S IV ] 10.5 m m; and [Ne II] 12.8 m m. These lines have been measured in a number ofH II regions in the inner Galaxy, and evidence for abundances elevated by factors of two or three have beenfound for the Galactic Center and for H II regions in the 5 kpc “ring” region (Pipher et al., 1984).However, even these mid-infrared lines suffer somewhat from extinction. In particular, the [Ar III ] and[S IV ] lines fall in the middle of the strong 10 m m silicate absorption feature, where the optical depth iscomparable to that in the near-infrared. Another approach to studying abundances in the inner galaxy isto make use of the fine-structure lines of [O III ] 52, 88 m m and [N III] 57 m m. By a happy coincidence,these lines from the abundant and (presumably) usually co-extensive O ++ and N ++ ions fall close togetherin wavelength and have fairly similar dependences on the electron density. The line emissivities are alsoessentially independent of the electron temperature. Measurements of these three lines therefore yield arelatively accurate value for the N/O ratio. A survey of about a dozen galactic H II regions in these linesyielded strong evidence that N/O in the Galactic Center and 5 kpc “ring” is elevated by a factor of 2 or 3 ascompared to the solar neighborhood. There remain some unsettled questions regarding N/O determinationsfrom the far-infrared lines, including possible ionization structure effects in H II regions ionized by verycool stars and a systematic discrepancy between values derived from the infrared lines and those derivedoptically from [N II ]/[O II ] (Rubin et al., 1988).More recent observations of IR fine-structure lines of the [S III ] 19 m m, [O III ] 52 and 88 m m, and [N III ]57 m m in compact H II Galactic regions have found abundance gradients of the form [S/H] = (-4.45 ± ± ± ± The recognition of significant variations in the gas composition within and among galaxies, along withparallel results on the stellar populations, inspired the development of chemical evolution models whichattempt to explain these patterns. The so-called “simple model” postulates a closed system of gas and stars,which self-enriches in metals as generations of stars age, die, and seed the ambient gas in the heavy elements(Searle & Sargent, 1972). This model also makes the approximations that the stellar lifetimes and timescalefor complete mixing of nucleosynthetic products are negligible in comparison to the timescale on which themetallicity evolves (“instantaneous recycling”). The simple model makes a specific prediction regarding themetallicity and system properties: Z = y ln ( M total / M gas ) , (1.1)in this equation Z is the metal abundance, y is the fraction of the stellar mass converted to heavy elements(the yield ), and M total = M gas + M stars .Although this model is most appropriate for the low-mass galaxies, it can also be applied to large diskgalaxies if concentric radii are treated as independent zones. However, it does not explain the observed gra-dients, so modifications such as radial flows, matter exchange with an outside reservoir (infall and outflow),or a variable stellar initial mass function, have been proposed as modifications to the model (Matteucci &Francois, 1989; Dopita, 1990).The relative abundances of nitrogen and oxygen are of particular interest, since they are synthesizedin different astrophysical sites. Oxygen is synthesized in massive stars and distributed into the interstellarmedium by Type II supernovae, while the origin of nitrogen is more problematic. A distinction is frequentlymade between “primary” nucleosynthetic products, which can be synthesized directly from H and He inPopulation III stars, and “secondary” products, which require a “seed” heavy nucleus to be initially presentin the star where its synthesis occurs. By this definition, oxygen is a primary species. Nitrogen is secondarywhen made as a by-product of CNO-cycle hydrogen burning. According to the simple closed-box model,the abundance of a secondary species is quadratic, so that if N is secondary and O primary, then (N/H) (cid:181) (O/H) , or (N/O) (cid:181) (O/H). The N/O ratio does appear to approach this behavior, for H II regions withmoderately high O/H values in M 101 (Torres-Peimbert et al., 1989) and in the Milky Way. However, belowa certain values of O/H, it appears that N/O is constant; these low-metallicity H II regions occur mostlyin low-mass galaxies. Thus, it is becoming clear that nitrogen is not purely a secondary nucleosyntheticproduct. Indeed, N may be produced within intermediate-mass stars by an effectively primary process, if Csynthesized within the star by the triple-alpha reaction is later subjected to the CN cycle. Nitrogen made bythis process would be primary, but there might be a time-delay in building up its abundance relative to thenuclear products of supernovae, because of the longer lifetimes of the source stars.The other elements measured in extragalactic H II regions, S, Ne, and Ar, are not likely to be dominatedby secondary processes. They might still, however, vary differently than oxygen, if they were produced instars of different mass ranges and the IFM varied or the timescales for enrichment differed substantially.There are known variations in the abundance ratios of certain elements. For example the fact that the iron-group is deficient relative to oxygen in Population II stars is thought to reflect an origin for the formerchiefly in Type I supernovae, which originate in long-lived progenitors, as opposed to synthesis of oxygenin massive stars and Type II supernovae.2 Chapter 1. IntroductionIn the context of chemical evolution models, Garnett (2002) studied the metallicity-luminosity andmetallicity-rotation speed correlations for spiral and irregular galaxies for a sample of spiral and irregu-lar galaxies having well-measured abundance profiles, distances, and rotation speeds. He finds that theO/H-V rot relation shows a change in slope at a rotation speed of about 125 km s − . At faster V rot , thereappears to be no relation between average metallicity and rotation speed. At lower V rot , the metallicitycorrelates with rotation speed. This change in behavior could be the result of increasing loss of metals fromthe smaller galaxies in supernova-driven winds. The idea was tested by looking at the variation in effectiveyield, derived from observed abundances and gas fractions assuming closed box chemical evolution. Theeffective yields derived for spiral and irregular galaxies increase by a factor of 10-20 from V rot ∼ − , asymptotically increasing to approximately constant y e f f for V rot ∼
150 km s − . The trend suggeststhat galaxies with V rot ∼ − may lose a large fraction of their supernova ejecta, while galaxiesabove this value tend to retain metals. The determination of effective yields as function of galactic radiusand its interpretation stands as one of the main studies in order to discriminate among different physicaleffects which may affect the chemical evolution of a galaxy. As described in this chapter, the study of chemical abundances has undergone a remarkable developmentin the last decades thanks mostly to important observational efforts that have focused on the derivation ofphysical and chemical properties of emission line H II regions in galaxies by spectroscopic techniques. Themain motivation common to all of these observations is to use the chemical information as one of the meansat our disposal to understand the physical processes at play in the formation and evolution of galaxies in theuniverse.Hitherto, most spectroscopic studies in nearby objects have been limited by the number of objects sam-pled, the number of H II regions observed and the coverage of these regions within the galaxy surface. Inorder to increase significantly the number of H II regions sampled in any given galaxy we require the com-bination of high quality multi-wavelength data and wide field spectroscopy. The advent of multi-objectand integral field spectrometers with large field of view now offer us the opportunity to undertake a newgeneration of observations, based on samples of scores to hundreds of H II regions and full 2-dimensional(2D) coverage. These sort of data would enable to test, confirm and extent the previous body of results fromsmall sample studies, while at the same time open up a new frontier of studying the 2D metallicity structureof disks and the intrinsic dispersion in metallicity, or to test and strengthen the diagnostic methods that areused to measure the H II region abundances in galaxies, among other issues.The scientific core of this dissertation is based on an observational project conceived to tackle the prob-lem of the 2D spectroscopic coverage of the whole galaxy surface. New techniques in imaging spectroscopy(or integral field spectroscopy, IFS) provide a powerful tool for studying the small and intermediate scale-size variation in line emission and stellar continuum in nearby well-resolved galaxies. We designed a projectto take advantage of these new observational techniques in order to assemble a unique spectroscopic sam-ple from which we could study, with unprecedented detail, the star formation and gas chemistry acrossthe surface of a galaxy. The observations consist of Integral Field Unit (IFU) 2D spectroscopic mosaicsof a representative sample of nearby galaxies ( D <
100 Mpc) with a projected angular size of less than10 arcmin. The mosaics were constructed using the unique instrumental capabilities of the Postdam MultiAperture Spectrograph, PMAS (Roth et al., 2005) in the PPAK mode (Verheijen et al., 2004; Kelz & Roth,.5Goals ofthis dissertation 132006) at the German-Hispanic Astronomical Centre at Calar Alto (CAHA), Spain. The PMAS fibre PAcK(PPAK) is one of the world’s widest integral field unit with a field-of-view (FOV) of 74 ×
65 arcseconds thatprovides a semi-contiguous regular sampling of extended astronomical objects. This project represents thefirst attempt to obtain 2D spectra of the whole surface of a galaxy in the nearby universe. The spectroscopicmosaicing comprises more than 50 000 spectra in the optical wavelength range.This observational project was devised as a scientific international consortium, the members are world-leading experts in their respective fields, including star formation and chemical abundances of galaxies,active galactic nuclei, multiwavelength observations of emission line regions and 2D spectroscopy. Theproject was entitled: the P PAK I FS N earby G alaxies S urvey, or PINGS . The P.I. of this project is Prof.Robert C. Kennicutt Jr. at the Institute of Astronomy, University of Cambridge. The collaborators of theconsortium are: Dr. ´Angeles D´ıaz at the Universidad Aut´onoma de Madrid, Spain; Dr. Anna Pasquali atthe Max-Plank Institut f¨ur Astronomie in Heidelberg, Germany; Dr. Sebasti´an S. S´anchez at CAHA, Spain;Benjamin Johnson and Caina Hao at the University of Cambridge, UK.The primary scientific objectives of this dissertation are to use the PINGS observations to obtain pixel-resolved emission-line maps across the disks of the galaxies to study the 2D abundance distribution and oncharacterising the relations between these abundance properties and the physical properties of the parentgalaxies. By targeting virtually every H II region in the galaxies, as a consequence of the nearly completespatial coverage of the IFUs, we are able to test for the first time the systematic dependences of the strong-line abundances on the size, luminosity, surface brightness, and other properties of the H II regions. Inthat respect, the PINGS observations and the subsequent analysis represent a leading leap in the studyof the chemical abundances and the global properties of galaxies, information which is most relevant forinterpreting observations at all redshift sources accessible with the current technology. The structure of this thesis is as follows: In §
2, I discuss the importance of Integral Field Spectroscopy (IFS)in astrophysics, including an explanation of the technique with their advantages and pitfalls, a description ofthe available instrumentation in the world by the time this project was envisaged, and the selection criteriaof the telescope-instrument chosen for this project. I also include a brief review to previous works that haveattempted to obtain 3D chemical abundance information in galaxies. In §
3, I present the sample of galaxiesand the selection criteria followed according to the scientific objectives established for this dissertation. Thischapter also includes a full description of the logistics and observations, explaining the telescope set-up andthe particular observing technique adopted for this project. In §
4, I explain the reduction process of the IFSraw data, the additional corrections implemented in this project and the improvements with respect to pre-vious pipelines, particularly regarding the flux calibration and the sky-subtraction. The possible sources oferrors and uncertainties are addressed, together with an explanation of the techniques implemented to min-imise them. In §
5, I present the integrated spectra of the PINGS sample, obtained by co-adding the spectrafrom their corresponding mosaics. Comparisons with previously published data are included. An analysisof the ionized gas component is performed, together with the techniques and methodologies implemented inorder to derive the physical parameters of the integrated gas-phase of the galaxies sample. In §
6, I presenta complete 2D spectroscopic study for a selected number of galaxies from the sample. Selected H II regionspreviously observed are compared with spectra extracted from the PINGS sample. A set of emission linemaps calculated from each galaxy is presented, including a quantitative description of the 2D distribution of4 Chapter 1. Introductionthe physical properties inferred from them. Then, a detailed, spatially-resolved spectroscopic analysis of theselected galaxies is performed, based on different spectral samples extracted from the full IFS mosaics ofthe galaxies. Several diagnostic diagrams and the state-of-the-art abundance diagnostic techniques are usedto obtain the 2D distribution of the physical properties and chemical abundances of the selected sample.Finally, in § HES IS CONTENT