Optical Spectroscopy of Halpha Filaments in Cool Core Clusters: Kinematics, Reddening, and Sources of Ionization
OOptical Spectroscopy of H α Filaments in Cool Core Clusters:Kinematics, Reddening and Sources of Ionization
Michael McDonald , , Sylvain Veilleux , and David S. N. Rupke , ABSTRACT
We have obtained deep, high spatial and spectral resolution, long-slit spectraof the H α nebulae in the cool cores of 9 galaxy clusters. This sample provides awealth of information on the ionization state, kinematics, and reddening of thewarm gas in the cool cores of galaxy clusters. We find evidence for only smallamounts of reddening in the extended, line-emitting filaments, with the majorityof filaments having E ( B − V ) < .
2. We find, in agreement with previous works,that the optical emission in cool core clusters has elevated low-ionization lineratios. The combination of [O
III ]/H β , [N II ]/H α , [S II ]/H α , and [O I ]/H α allowus to rule out collisional ionization by cosmic rays, thermal conduction, andphotoionization by ICM X-rays and AGN as strong contributors to the ionizationin the bulk of the optical line-emitting gas in both the nuclei and filaments. Thedata are adequately described by a composite model of slow shocks and starformation. This model is further supported by an observed correlation betweenthe linewidths and low ionization line ratios which becomes stronger in systemswith more modest star formation activity based on far ultraviolet observations.We find that the more extended, narrow filaments tend to have shallower velocitygradients and narrower linewidths than the compact filamentary complexes. Weconfirm that the widths of the emission lines decrease with radius, from FWHM ∼
600 km s − in the nuclei to FWHM ∼
100 km s − in the most extended filaments.The variation of linewidth with radius is vastly different than what is measuredfrom stellar absorption lines in a typical giant elliptical galaxy, suggesting thatthe velocity width of the warm gas may in fact be linked to ICM turbulence and, Kavli Institute for Astrophysics and Space Research, MIT, Cambridge, MA 02139, USA Department of Astronomy, University of Maryland, College Park, MD 20742, USA Department of Physics, Rhodes College, Memphis, TN 38112, USA Email: [email protected] Email: [email protected] Email: [email protected] a r X i v : . [ a s t r o - ph . C O ] N ov r <
10 kpc) of several systems the warm gas shows kinematicsignatures consistent with rotation, consistent with earlier work. We find that thekinematics of the most extended filaments in this sample are broadly consistentwith both infall and outflow, and recommend further studies linking the warmgas kinematics to both radio and X-ray maps in order to further understand theobserved kinematics.
Subject headings: galaxies: cooling flows – galaxies: clusters – galaxies: groups –galaxies: elliptical and lenticular, cD – galaxies: active – ISM: jets and outflows
1. Introduction
The dense core of a galaxy cluster represents a unique environment, where the hotintracluster medium (ICM) is cooling most rapidly, feedback from the central active galacticnucleus (AGN) is most effective, and the brightest cluster galaxy (BCG) dominates themass. This represents one of the few places in the Universe where large-scale cooling andfeedback processes can be readily observed. Unlike isolated massive galaxies, where some ofthe energy injected into the interstellar medium (ISM) from the AGN often escapes into thelow-density intergalactic medium (IGM), the denser ICM in cluster cores retains an imprintof this feedback in the form of bubbles (e.g., Churazov et al. 2001; Reynolds et al. 2005;Revaz et al. 2008) or ripples (e.g., Perseus A; Fabian et al. 2003). Similarly, the accretionof hot gas from the IGM onto massive galaxies is challenging to observe due to the very lowdensities, while such phenomena have been studied in depth for decades in galaxy clusters.The so-called “cooling flows” in galaxy clusters, which were once thought to be massiveflows of cool gas on the order of 100–1000 M (cid:12) yr − (see review by Fabian 1994), are nowunderstood to be considerably less massive, depositing on the order of 1–10 M (cid:12) yr − of coolgas onto the BCG (e.g., Voigt & Fabian 2004; Bregman et al. 2006). It is assumed thatsome form of feedback (e.g., AGN, gas sloshing, conduction) prevents the dense ICM fromcatastrophically cooling, allowing only a trickle of cool gas to accrete onto the BCG.There is a considerable amount of support for the “reduced cooling flow” model (Voigt& Fabian 2004) from observations at a variety of wavelengths. Cool core clusters (galaxyclusters for which the central ICM cooling time is much less than the age of the Universe)are well known to have an abundance of multi-phase gas in the central ∼
100 kpc whichhas been observed in O VI (e.g., Bregman et al. 2001; Oegerle et al. 2001; Bregman et al.2006), H α (e.g., Hu et al. 1985; Johnstone et al. 1987; Heckman et al. 1989; Crawfordet al. 1999; Jaffe et al. 2005; Edwards et al. 2007; Hatch et al. 2007; McDonald et al. 2010, 3 –2011a), and H (e.g., Edge 2001; Salom´e & Combes 2003; Jaffe et al. 2005; Lim et al. 2008;Donahue et al. 2011). Additionally, evidence for star formation has been observed in the UV(e.g., McNamara & O’Connell 1989; Rafferty et al. 2006; McDonald & Veilleux 2009; Hickset al. 2010; McDonald et al. 2011b), optical (e.g., Allen 1995; Cardiel et al. 1995; Crawfordet al. 1999; Edwards et al. 2007; Bildfell et al. 2008), and mid IR (e.g., Hansen et al. 2000;Egami et al. 2006; O’Dea et al. 2008; Quillen et al. 2008, hereafter MIR). In McDonaldet al. (2011b), we showed that the star formation rates inferred from far-UV and mid-IRdata suggest an efficiency of ∼
15% in converting the cooling ICM into stars, consistentwith the global baryon fraction in stars. This star formation efficiency estimate, based onfar-UV, H α , and mid-IR data, is consistent with earlier work based on mid-IR star formationestimates by Egami et al. (2006) and O’Dea et al. (2008). This overwhelming evidence forcooling byproducts suggests that cooling flows are occurring, but on a reduced scale thanwas initially predicted.In McDonald et al. (2010, 2011a) we presented a sample of 33 galaxy groups and clus-ters with deep, high-spatial-resolution H α imaging from the Maryland-Magellan TunableFilter (MMTF ; Veilleux et al. 2010). We reported on several new systems with extended,filamentary H α emission, and produced higher-quality H α maps for some previously knownsystems. By combining deep, archival Chandra X-ray Observatory (CXO) data with thesenew H α images, we showed that the extended warm gas (H α ) is spatially coincident with theasymmetric, rapidly-cooling ICM for an ensemble of optically-emitting nebulae and that theH α emission is always confined within the ICM cooling radius. This, taken together with anumber of previously known correlations between the warm and hot phases (see review byFabian 1994), showed that the H α emission is intimately linked to the X-ray cooling flow inboth quantity and morphology. However, while we were able to determine the origin of thiscool gas (cooling out of the hot ICM), we were unable to constrain the dominant ionizationprocesses which produce the high H α fluxes. In an attempt to shed light on this remainingmystery, we have acquired deep, high-resolution spectra of multiple filaments and BCG nu-clei in a sample of cool core clusters drawn from our previous works (McDonald et al. 2010,2011a,b). We present these data in §
2, describing their collection, reduction, and analysis.In § § § = 73 km s − Mpc − , Ω matter = 0.27, Ω vacuum = 0.73.
2. Data Collection and Analysis
Name z E(B-V) Source Slit PAs H α Ref.(1) (2) (3) (4) (5) (6)Abell 0478 0.0881 0.517 Keck – LRIS 356. ◦
4, 6. ◦ ◦
4, 40. ◦ ◦
0, 72. ◦ ◦
0, 79. ◦ ◦
4, 344. ◦ ◦
1, 108. ◦ ◦ ◦ ,5 163. ◦ ◦
0, 161. ◦ α data.In order to investigate the various properties of the optically-emitting filaments in coolcores, we performed long-slit spectroscopy on a sample of 9 BCGs. This sample was chosento include the most extended and luminous systems at H α from our larger sample of 36clusters with deep H α imaging from the Maryland-Magellan Tunable Filter (Veilleux et al.2010) as described in McDonald et al. (2010, 2011a,b) (Table 1). Long-slit spectra wereobtained over the span of 2 years at the Magellan and Keck telescopes, using the Inamori-Magellan Areal Camera & Spectrograph (IMACS; Dressler et al. 2011) and Low ResolutionImaging Spectrometer (LRIS; Oke et al. 1995), respectively. The setup we used allowedfor a broad wavelength coverage, allowing full spectral coverage from ∼ ∼ ∼
75 and 145 km s − for IMACS and LRIS@ 7000˚A, respectively). For each cluster, two slits were aligned along the filaments in anattempt to maximize the coverage of the optically-emitting gas. Additionally, one of the twoslits was always forced to pass through the BCG nucleus. In the case of Abell 2597 only a 5 –single slit position was acquired due to time constraints, and in NGC 4325 one of the twoslits is slightly offset from the optimal position. The chosen slit positions are shown in Fig.1. In total, 17 optical spectra of 9 cluster cores were obtained. These spectra were reducedusing standard IRAF ( http://iraf.net ) packages for the reduction of long-slit spectra.This procedure involved removing the zero-point bias from and flat-fielding each exposure(CCDPROC), masking cosmic rays (LA Cosmic; ), combining similar exposures (IMCOMBINE), removing spatial and spectral dis-tortions (TRANSFORM), removing sky lines (XVista; http://ganymede.nmsu.edu/holtz/xvista/ ), and calibrating the spectra in both wavelength (IDENTIFY, FITCOORDS) andintensity (STANDARD, SENSFUNC, CALIBRATE). The wavelength calibration was de-rived from spectra of helium, neon, and argon arc lamps, while the photometric calibrationwas based on the spectrophotometric flux standards EG274, LTT3218, and G191-B2B. Spec-tra were visually inspected for contamination of the emission lines from the O atmosphericabsorption features around 6870˚A and 7620˚A.From each spectrum, the redshift and velocity dispersion of the stars at the center ofthe BCG were measured using the Na D lines at 5890˚A and 5896˚A and the Mg I line at5178˚A wherever available. Next, the H α and [N II ] lines, which do not require deconvolutiondue to the high spectral resolution, were simultaneously fit using a combination of threeGaussians. It was assumed that all three lines had the same velocity dispersion and redshift,which yields 4 free parameters: v, σ v , F Hα , F [ N II ] . In the few cases where multiple velocitycomponents were visible, we fit two Gaussians to each line. The continuum was fit using athree-parameter model, which consisted of a linear component (2 parameters) and an H α stellar absorption feature with the width and redshift fixed at the value measured from theNa D and Mg I lines. All spectra were visually inspected for contamination of the emissionlines from the O atmospheric features at 6870˚A and 7620˚A. Of our 9 targets, only Abell 1644is redshifted such that the H α +[N II ] lines are affected by O absorption. For this system,we clipped the spectrum from 6870–6876˚A, which removed the strongly-absorbed part of thespectrum while leaving the majority of the emission lines intact. We note that, while the H α line fluxes in this system are likely still contaminated and should be taken as lower limits,they do not appear as outliers in any of the observed trends and have H β fluxes and H α FWHM consistent with the low observed flux, as we will show.Following the fits to the H α +[N II ] lines, the H β , [O III ], [O I ], and [S II ] lines wereindividually isolated and fit with Gaussians. For each of these fits the redshift and velocitydispersion were fixed to the values for H α , while the line intensity and continuum zeropointand slope were allowed to vary. Additionally, an H β stellar absorption feature was included 6 –Fig. 1.— Slit positions for all 9 cluster cores in our sample overlaid on H α images fromMcDonald et al. (2010, 2011a,b). The length of the slits have been artificially truncated forthis plot to cover only optically-emitting regions. In all cases we have attempted to maximizethe overlap between these slits and the H α maps while forcing one slit to pass through theBCG nucleus. In the case of Abell 2597 we were only able to obtain one pointing due totime constraints. In NGC 4325 one of the slits is slightly offset from the ideal position dueto a pointing error. 7 –with a variable amount of absorption. Finally, a visual inspection of all 5 fitting regions wasperformed for each spectrum and spurious fits were re-run with more specific input parame-ters, allowing for a more reliable fit. All measured line intensities were further corrected forGalactic extinction following Cardelli et al. (1989) using reddening estimates from Schlegelet al. (1998).
3. Results
These data, described in full in Appendix A, provide a wealth of information about thewarm gas kinematics, ionization state, and amount of reddening in the cool cores of galaxyclusters. In this section, we first address the results of our spectroscopic analysis in thecontext of these three major topics, and then follow with a discussion of the implications thatthese results have on our current understanding of the generation of these optical filaments.
The use of deep, high spatial resolution, narrow-band H α imaging to dictate the ori-entation of our long-slit spectroscopy allows us to characterize the properties of the opticalemission in two spatial dimensions without the field-of-view limitations of 3-D spectroscopy.The broad wavelength coverage of these spectra allows the comparison of various opticalemission lines, specifically the key diagnostic line ratios [N II ] λ α , [O I ] λ α ,[S II ] λλ α , and [O III ] λ β . This combination of line ratios is tradi-tionally used to identify the source of ionization in emission-line regions (Baldwin et al.1981; Veilleux & Osterbrock 1987; Kewley et al. 2006). We show in Figure 2 the first ofthese ratios, [N II ]/H α , which is often used to differentiate between star-forming regionsand AGN, overlaid on H α maps of individual cluster cores. Only 2/9 clusters (Abell 1644,Abell 2052) have [N II ]/H α strongly peaked on the nucleus, where we would expect AGN,if present, to dominate the ionization. Roughly 2/3 of systems have low ( (cid:46) .
5) [N II ]/H α everywhere, while the remaining 1/3 of systems have high ( (cid:38) .
5) [N II ]/H α everywhere.Interestingly, this is the same fraction of cool cores that appear to be star forming (McDon-ald et al. 2011b). For systems with both far-UV imaging from McDonald & Veilleux (2009)and McDonald et al. (2011b) and optical spectroscopy, we find a one-to-one correspondencebetween star-forming (Abell 1795, Abell 2597) and shock-heated (Abell 1644, Abell 2052)systems based on [N II ]/H α and far-UV/H α flux ratios. The systems with the longest fil-aments (Abell 1795, Abell 2597, Sersic 159-03) have the lowest overall [N II ]/H α ratios,with the value of the ratio decreasing slightly at larger radius. However, these [N II ]/H α II ] λ α intensity ratio. The combination ofnarrow-band H α imaging with long-slit spectroscopy allows us to create pseudo-2D spectra.In all panels, both the point size and color correspond to the magnitude of the line ratio, asdescribed by the legend in the upper left. The cluster or central galaxy name is shown in theupper right, while the physical scale of the image is shown via a 10kpc scale bar in the lowerright. The white cross represents the center of the optical (stellar continuum) emission. 9 –Fig. 3.— Key diagnostic diagrams involving the reddening-insensitive line ratios [N II ] λ α , [O I ] λ α , [S II ] λλ α , and [O III ] λ β for filaments(stars) and nuclei (circles) in each of our 9 systems. The grey points are galaxies from theSloan Digital Sky Survey (Kewley et al. 2006) and show a clear separation into H II regionsand AGN. The red line represents the extreme starburst limit and separates HII regionsfrom AGN, while the blue line separates AGN into Seyfert and LINER classes (Kewleyet al. 2006). Typical errorbars are shown in the bottom left corner of each panel. We notethat, in general, the data for optical emission in cool core clusters tend towards elevatedlow-ionization lines. 10 –ratios are significantly elevated above what one would expect for ongoing star formation( ∼ . ∼ .
0; e.g., Kewley et al. 2006). Several prior studies have also found elevatedlow-ionization line ratios in the optical emission in cool core clusters (e.g., Voit & Donahue1997; Crawford et al. 1999). In order to understand these line ratios, we appeal to the fullsuite of line ratios first used by Veilleux & Osterbrock (1987).In Figure 3, we show the [O
III ]/H β line ratio as a function of [N II ]/H α , [O I ]/H α , and[S II ]/H α . In these plots we show data from the Sloan Digital Sky Survey (SDSS; Abazajianet al. 2009) for emission-line nuclei, showing the separate regions occupied by star-forminggalaxies, low-ionization nuclear emission-line regions (LINERs), and Seyfert galaxies. Wefind a systematic offset between the emission-line ratios in the warm filaments and the locusof points from the SDSS, towards lower [O III ]/H β and/or higher [N II ]/H α , [S II ]/H α , and[O I ]/H α . We note that the grey points represent nuclei , while much of our data includesa stronger contribution from diffuse, ionized regions. The fact that excess blue light (e.g.,Crawford et al. 1999; Edwards et al. 2007; Bildfell et al. 2008), UV emission (e.g., Raffertyet al. 2006; Hicks et al. 2010; McDonald et al. 2011b), mid-far IR emission (e.g., O’Deaet al. 2008), and molecular gas (e.g., Edge 2001; Salom´e & Combes 2003) are observedis strong evidence for star formation, but Figure 3 seems to indicate that the situation ismore complex. The [N II ]/H α and [S II ]/H α ratios in some filaments are consistent withthe extreme starburst limit (Kewley et al. 2006), however the [O I ]/H α ratio is far too highto be pure star formation. Thus, it appears that pure star formation, whether ongoing orburst-like, is unable to explain all of the observed optical line ratios in the cool cores ofgalaxy clusters. In § III ]/H β ,consistent with LINERs. This classification is consistent with the relative weak luminosityof nuclei in cool core BCGs in both the optical and X-ray and with the narrow emission lines(Figure 10). However, only three nuclei (Abell 0496, Abell 1644, and Abell 2052) are clearlyoffset from the filaments in Figure 3. The remaining six nuclei are likely undergoing thesame ionization processes as the filaments and, thus, may not be AGN, or may be compositeobjects (Kewley et al. 2006), despite their identification as LINERs in Figure 3.The high spectral resolution of these data allows us to separate the [S II ] doublet anddetermine the [S II ] λ λ II ] λ λ − , consistent with earlierwork by Heckman et al. (1989). The more extended gas in the filaments appears to haveoverall lower densities, although we are unable to determine the average density due to theinability of the [S II ] λ λ ∼
100 cm − . In general, thefilaments tend to have N e <
200 cm − , potentially reaching much lower than ∼
100 cm − atlarge radius.Fig. 4.— Distribution of the [S II ] λ λ − , the [S II ] λ λ II ]/H α line ratios to those from the literaturein order to assess the quality of these measurements. There is generally good agreementbetween this work and the literature, suggesting that our measured ratios are reliable towithin ∼ II ]/H α line ratios for this work to those taken from the literature.Blue and red points represent extended and nuclear emission, respectively. Point typescorrespond to different references: Heckman et al. (1989) – filled circles, Melnick et al. (1997)– filled squares, Jaffe et al. (2005) – filled triangles, Hatch et al. (2007) – stars. The dashedline represents equality, while the dotted lines represent a scatter of ∼ II ]and H α lines.clusters. Furthermore, we find that the optical line ratios measured in Centaurus A (Farageet al. 2010) agree with both Perseus A and our sample. Overall, the filaments in cool corestend to have elevated low-ionization ratios, with the [O I ]/H α ratio being very narrowlypeaked at ∼ In Figure 7 we show H α maps for each cluster core, with the line-of-sight velocity mea-surements overlaid. These velocities represent the line-of-sight velocity difference betweenthe warm gas and the central (BCG) stellar component. We find that, in general, sys-tems with extended emission tend to have relatively smooth velocity fields (e.g., Abell 0496,Abell 1795, Abell 2597, Sersic 159-03), while compact systems and nuclei tend to have rota-tion signatures (e.g. Abell 0780, Abell 1644, Abell 0478). This observation is consistent with 13 –Fig. 6.— Distribution of reddening-insensitive optical line ratios measured in Perseus A (toprow; Hatch et al. 2006), nuclei (middle row) and filaments (bottom row) for the 9 clusters inour sample. Note the overlap between the measured line ratios in Perseus A and in a typicalfilament.the work of Baum et al. (1990) and Baum et al. (1992) who found that emission-line nebulaein radio galaxies can be divided into three kinematic classes: rotators, calm nonrotators,and violent nonrotators. The systems which are classified as nonrotators (both calm andviolent) tend to have relatively constant velocity fields and resemble the filaments in Figure7, while the rotators have strongly varying velocity fields, much like the nuclei and compactsystems in Figure 7. This is more easily seen in Figure 8, which shows the line-of-sightvelocity fields. In the central regions ( <
10 kpc) the velocity fields in many systems showevidence for rotation, with a characteristic shape reminiscent of a rotating disk. At largeradius ( >
10 kpc), the velocity fields of the filaments appear to be relatively flat with typicalline-of-sight velocities of ∼
300 km s − relative to the BCG. We do not find a correspondencebetween the kinematic class (rotator or non-rotator) as defined by Baum et al. (1990, 1992)and the ionization class (star-forming or shock-heated) described in § § < v/σ (Baum et al. 1992). Negative and positive radii correspond topositions along the slit with RA less than and greater than the BCG center, respectively. 16 –Fig. 9.— Similar to Figure 2 but now showing the velocity width. In nearly all systems thevelocity dispersion peaks at the position of the optical nucleus. In the more extended systemsthere is a trend towards decreasing velocity dispersion with increasing radius, suggesting thatturbulence is larger in the cool cores. 17 –Fig. 10.— Left: Similar to figure 8, but now showing optical linewidth as a function ofradius for individual clusters. Right: Optical linewidth as a function of radius for all 9galaxy clusters. There is a strong correlation, such that the warm gas in the very center ofthe cluster is highly turbulent (FWHM ∼
400 km s − ) while the thin, extended filamentshave very narrow lines (FWHM <
200 km s − ). This trend suggests that highly-elongatedfilaments can only survive in regions with minimal turbulence. Shown in blue is the stellarabsorption line width in M87 from Murphy et al. (2011).In Figure 9 we show the velocity width of the filaments. In general, the velocity widthpeaks in the nucleus (FWHM ∼
400 km s − ) while, in the filaments, it is consistently lower(FWHM ∼ −
200 km s − ). The longest filaments (i.e., Abell 1795, Abell 2597, Sersic 159-03, NGC 4325) tend to have the narrowest emission lines. Curiously, Abell 2052, whichharbors the most radio-luminous central AGN in our sample has the lowest central velocitywidth. Overall, the velocity width correlates well with the amount of turbulence one wouldinfer from a visual inspection: long, thin filaments (e.g., Abell 1795, Abell 2597, Sersic 159-03) tend to have narrow emission lines, while short stubby filaments (e.g., Abell 0478) andregions with “disturbed” morphologies (e.g., Abell 0496, Abell 0780) tend to have broaderprofiles.Figure 10 shows the variation of the velocity width as a function of projected radiusfor each spatial element shown in Figure 9. There is an obvious correlation (Pearson R = − .
63) between the optical linewidth and radial extent of the filaments. This trend was alsoreported, albeit with lower significance, by Baum et al. (1990) and Baum (1992). This figure 18 –shows compelling evidence that the most extended optical filaments are also those whichare experiencing the least amount of turbulence. If we assume that the H α -emitting gasis a product of the ICM cooling (e.g., McDonald et al. 2010), this may tell us somethingabout the amount of turbulence in the ICM within the cooling radius. Beyond the inner10kpc, the optical linewidth is <
200 km s − , which is consistent with the observed upperlimits on the ICM turbulence in cool cores from the XMM Reflection Grating Spectrometer ( ∼
200 km s − ; Sanders et al. 2011). We note that the instrumental FWHM is ∼
75 km s − ,so our measurements are well above the lower limits.The fact that we see broader lines near the center of the cluster could also be due tothe higher likelihood of chance superposition of filaments at smaller radii, where the numberdensity of filaments is higher. While the kinematic data alone are unable to differentiatebetween a true increase in velocity dispersion and a projection of multiple narrow lines, weare able to offer evidence for the former by looking at the optical line ratios. In § II ]/H α ratio is typically peaked in the nucleus, while in § (cid:46)
30% in radial velocityand velocity dispersion, respectively. We note that the fractional error in relative velocities(i.e. velocity of gas with respect to the BCG) is much lower and is independent of ourabsolute velocity calibration. The larger deviations in the velocity dispersion measurementsare most likely a result of differing apertures between this work and previous studies of thesame sources. Overall, there is good agreement between our kinematic measurements andthose from previous studies.These data tell an interesting story about the kinematics of the warm gas in the coolcores of galaxy clusters. In general, we find that longer filaments are less turbulent thanshorter or morphologically peculiar filaments. The velocity width in the warm gas increaseswith decreasing radius, reaching a peak in the nucleus of ∼
400 km s − , and a minimumof ∼
100 km s − beyond 10 kpc. The velocity fields are typically well-ordered along thefilaments. We return to these results in § The amount of dust in the ICM is still uncertain. The abundance of relativistic particlesshould act to destroy dust grains on short timescales (Draine & Salpeter 1979). However,the fact that we observe star formation and, more importantly, mid–far IR emission (e.g.,O’Dea et al. 2008), suggests that some amount of dust may be shielded from destructiveprocesses. Our data provide new estimates on the amount of intrinsic reddening in the opticalfilaments and nuclei in the BCGs of cool core clusters – something that is currently lackingin the literature for more than a few systems (e.g., Hu et al. 1985; Crawford et al. 1999).In order to determine the amount of intrinsic reddening, we began with the H α /H β lineratio, corrected for Galactic extinction and underlying stellar absorption ( § α /H β > .
85 (corresponding to case B recombination) hadnon-zero intrinsic reddening and determine the intrinsic E ( B − V ) assuming a dust screen 20 –model and a standard extinction curve ( R v = 3 .
1; Cardelli et al. 1989). The reddeningestimates here assume a foreground screen model and, thus, are underestimated by a factorof [exp( τ ) − /τ if we assume that the line-emitting gas and dust are well-mixed.In Figure 12 we provide the 2-D distribution of reddening estimates based on ( E ( B − V )),corrected for Galactic extinction, for the 9 systems in our sample. Typical uncertainties inthese estimates, assuming that the ∼ Hβ dominates, is ∆ E ( B − V ) ∼ .
15. With the exception of a few outliers, the vast majority of systems appear to bemostly free of intrinsic reddening. The five most extended systems, Abell 1644, Abell 1795,Abell 2597, NGC 4325, and Sersic 159-03 all appear nearly reddening-free. There is a sig-nificant amount of reddening in the core of Abell 0780 (Hydra A), which is known to havea dust-lane at approximately the position of the reddening peak. Abell 0478, which hasthe highest Galactic E ( B − V ), seems to have a moderate amount of intrinsic reddeningeverywhere, but we cannot formally rule out a Galactic origin due to small fractional un-certainty on the large Galactic extinction value. Both Abell 0496 and Abell 2052 appear tohave legitimate, non-zero reddening in the filaments. In the case of Abell 2052, the reddenedfilaments (north of the BCG nucleus) appear to be shock-heated (McDonald et al. 2011b)based on the high H α /FUV ratio.We can also consider the distribution of reddening measurements in filaments and nuclei.Using the same regions as in Figure 3, we take the emission-weighted average E ( B − V ) ineach region. The distribution of these values is shown in Figure 13. For the nuclei there isa relatively broad distribution in E ( B − V ) from 0.0–0.7, while in filaments the distributionpeaks at E ( B − V ) = 0 with a broad tail out to E ( B − V ) = 0 .
6. We find that ∼
70% offilaments have E ( B − V ) < .
2, contrary to previous estimates based on randomly-orientedslits (e.g., Crawford et al. 1999) which found considerable amounts of optical reddening incool core clusters. However, previous measurements contained only the nuclear emissionand, thus, a comparison to our in-filament reddening estimates is unfair. We find goodagreement between this work and Crawford et al. (1999) for the distribution of reddeningin nuclei, as shown in Figure 13. The consistently small amount of reddening derived inthe H α filaments appears to be inconsistent with the evidence in the UV and mid-infraredof on-going star formation (e.g., Rafferty et al. 2006; O’Dea et al. 2008; Hicks et al. 2010;McDonald et al. 2011b), but is consistent with previous findings for the warm gas in PerseusA (E(B-V) ∼ α /H β = 2 .
85 as the dust-free limit and donot change dramatically if we instead choose H α /H β = 3 .
1, which reflects the conditions inshock-heated gas. 21 –Fig. 12.— Similar to Figure 2 but now showing the intrinsic reddening, E(B-V), corrected forGalactic extinction. We note that many of the regions with locally high E(B-V) are coincidentwith low H α surface brightness (e.g., Abell 0496, Abell 1795, Sersic 159-03) suggesting thatthey are more uncertain. 22 –Fig. 13.— Distribution of optical reddening, E ( B − V ), in the nuclei (upper) and filaments(lower) of 9 cool core clusters. The amount of reddening in nuclei is relatively flat overthe range 0 < E ( B − V ) < .
4, while, in the filaments, the distribution peaks sharply at E ( B − V ) = 0, with a tail out to E ( B − V ) ∼ .
6. The distribution of E ( B − V ) in thenuclei matches well with that found by Crawford et al. (1999) (dotted line). The fact thatthe majority of filaments have little-to-no reddening suggests that dust does not survive longin the intracluster medium. The objects with E ( B − V ) > . τ ) − /τ if we assume thatthe line-emitting gas and dust are well-mixed.
4. Discussion
The data presented in § There are several possible sources of ionization in the cool cores of galaxy clusters. Themost popular ideas have been: (a) the central AGN, (b) young stellar populations, (c) X-rays from the ICM, (d) heat conduction from the ICM to the cold filament, (e) shocks andturbulent mixing layers, and (f) collisional heating by cosmic rays. The lack of a stronggradient in [N II ]/H α with radius (Figure 2) and the low [O III ]/H β ratio in the filamentssuggests that AGN do not contribute significantly to the ionization outside of the nucleus.The relative weakness of high-ionization lines (i.e., [O III ]) suggests that ionization by ICMX-rays is also a small contributor. While scenarios (a) and (c) are relatively easy to rule outin the filaments, the remaining scenarios require a more quantitative approach.In Figure 14 we show the same data as in Figure 3, but now include model expectationsfor star formation, cooling plasma, shocks, conduction, and collisional ionization by cosmicrays. The model grids representing photoionization by young stars (upper panels; Kewleyet al. 2001) show partial overlap with the data for filaments in the [N II ]/H α and [S II ]/H α panels. However, these models fail to produce adequately high [O I ]/H α ratios, as noted byearlier studies (e.g., Crawford et al. 1999). The best match between the model and data isachieved if we use a model cloud with roughly solar metallicity and a low total ionizationparameter (U). The location of this grid would move slightly if the IMF was altered, but noreasonable combination of ionization parameter, metallicity, and IMF would reproduce theobserved [O I ]/H α ratios.In the second row of Figure 14, we show the expectation for condensing intraclustergas, originating at 10 K (Voit et al. 1994). These models assume that the intracluster gasat
T > K photoionizes the 10 K gas as it cools radiatively. In choosing the models toplot we make the assumption that the cool gas is ionization-bounded. This assumption isbased on the fact that we observe strong [O I ] and [S II ] lines – species that are typicallyobserved beyond the classical H II region (Hatch et al. 2006). These models reproduce wellthe observed [N II ]/H α and [O III ]/H α ratios observed in the filaments, but tend to slightlyover-predict the [S II ]/H α and [O I ]/H α ratios. Interestingly, these models err in the oppositedirection as the models for star formation, suggesting that a combination of the two mayprovide an adequate fit to the data. An important test for this model is whether the observedcooling rates can produce the high optical line luminosities that we observe. In Donahue & 24 –Fig. 14.— Similar to Figure 3, but with model predictions overlaid. From top to bottom:photoionization from young stars (lines of constant ionization parameter and metallicity;Kewley et al. 2001), self-ionization from a condensing plasma (lines of constant ionizationparameter and metallicity; Voit et al. 1994), fast shocks (lines of constant speed and magneticfield strength; Allen et al. 2008), conduction (points correspond to different initial conditions;Boehringer & Fabian 1989), and collisional ionization by cosmic rays (Ferland et al. 2009).In all panels, the point/color types are consistent with those in Figure 3 and the red dashedline is the extreme starburst limit of Kewley et al. (2006). 25 –Fig. 14.— Continued.Voit (1991), the total H α luminosity due to the condensing ISM is given as: (cid:18) L Hα erg s − (cid:19) = (cid:32) ˙ M Hα
16 M (cid:12) yr − (cid:33) ( T f c (cid:15) ) (1)where T is the maximum temperature in units of 10 K, f c is the fraction of the coolingradiation incident upon the photoionized clouds, and (cid:15) is the fraction of incident radiationthat reemerges in the H α line normalized to 3%. We use (cid:15) = 0 .
77, from Voit et al. (1994).In Figure 15 we show the H α luminosity as a function of the X-ray-derived cooling ratesfor 14 clusters (red) from McDonald et al. (2010) and 7 groups (blue) from McDonald et al.(2011a). Even if we make the unrealistic assumption that 100% of the cooling luminosity isincident upon the cool clouds ( f c = 1), the majority of systems are still much more luminousat H α than predicted. Under these extreme conditions, the most luminous systems areconsistent with still only <
20% of their H α emission coming from the condensing ICM, inagreement with earlier results by Voit & Donahue (1990) using the same methods. Thus,while the model of a self-radiating cooling flow is promising in terms of the predicted lineratios, the total fluxes are too low in the context of more recent, spectroscopically-derivedX-ray cooling rates. However, the fact that we do measure non-zero cooling rates in the ICMmeans that self-radiation due to the cooling ICM is a contributing process and should notbe ignored – a point we will return to later.In the third row of Figure 14, we show model expectations for shocks of various speedsand magnetic field strengths from Allen et al. (2008). In general, the shock models coverthe same range in [N II ]/H α , [S II ]/H α , and [O I ]/H α as the data, but slightly over-predict[O III ]/H β . The low-ionization ratios are best matched in the filaments by slow shocks 26 –Fig. 15.— Total H α luminosity versus X-ray-derived cooling rates from McDonald et al.(2010) and McDonald et al. (2011a). The red and blue points correspond to galaxy clustersand groups, respectively. The dashed lines show the predicted H α luminosity assuming thatthe flux is due to self-ionization by a condensing hot plasma, for covering factors of 25%and 100%. The dotted line is five times the expected H α luminosity for a covering factorof 100%. For the most extended and luminous filaments, this process contributes, at most,20% of the H α flux, assuming an unrealistically high covering factor.( v ∼ −
400 km s − ), and in the nuclei by fast shocks ( v ∼
800 km s − ). This makesqualitative sense, since the line widths are considerably broader in the nucleus than in thefilaments. To further investigate the role of shocks, we consider the velocity dispersion asa function of the [N II ]/H α , [S II ]/H α , and [O I ]/H α ratios in Figure 16. We find that thesystems with the weakest low-ionization line ratios have velocity dispersions consistent withthose found for luminous and ultraluminous infrared galaxies (hereafter LIRGs and ULIRGs,respectively), which are likely experiencing a mix of shocks and star formation (Veilleux et al.1995, 1999; Monreal-Ibero et al. 2010). The systems which have low ionization line ratiosconsiderably larger than those found in ULIRGs and LIRGs were identified earlier as non starforming based on the optical line ratios and far-UV emission. In the right panels of Figure16, we show that these non star forming systems have [N II ]/H α and [S II ]/H α ratios which 27 –are more strongly dependent on the velocity dispersion than their star forming counterparts.One possible interpretation of this result is that the filaments and nuclei are ionized by amix of shocks and star formation. In this scenario, the two lines in Figure 16 represent twoextremes: the shock-dominated case (steep line) and star formation-dominated case (shallowline). We will further investigate this scenario in the next section.In the fourth row of Figure 14 we show the expectation for heat conduction along theboundary between the hot ICM and the cool filaments from Boehringer & Fabian (1989).This mechanism produces either too high [O III ]/H α ratio or too low [N II ], [S II ], and [O I ]ratios, depending on the choice of conditions. While the presence of soft X-ray emissioncoincident with optical filaments is evidence that conduction may be occurring (e.g., Nipoti& Binney 2004), it cannot be the dominant source of ionization of the optical filaments onthe basis of the observed optical line ratios.Finally, the bottom panels of Figure 14 shows the predicted line ratios for collisionalionization by cosmic rays (Ferland et al. 2009). While this model accurately predicts the[S II ]/H α and [O I ]/H α ratios, it underpredicts the [N II ]/H α and [O III ]/H β ratios by 1and 3 orders of magnitude, respectively. In the absence of an additional ionization sourcewhich can produce strong [O III ] and [N II ] emission while producing very little H α , H β ,[S II ], and [O I ], these large discrepancies effectively rule out cosmic rays as a stronglycontributing ionization mechanism. However, we point out that, while they do not appearto be a sufficient ionization mechanism for the cool gas, cosmic rays may still be importantin providing energy to the ICM, heating H (Ferland et al. 2009; Donahue et al. 2011), andhelping to prevent massive cooling flows (e.g., Mathews 2009).It is clear from Figures 14–16 that, individually, none of the models we listed at thebeginning of this section can simultaneously explain all optical line ratios, their observeddependence on linewidths, and the total H α luminosity. However, there is a very real pos-sibility that these filaments are, in fact, ionized by a combination of processes. Below, wediscuss two such composite models which offer promising results. There is compelling evidence for both star formation and cooling flows in the cool coresof many galaxy clusters. Modest cooling flows are observed in the X-rays (e.g., Petersonet al. 2003; Voigt & Fabian 2004) and EUV (e.g., Oegerle et al. 2001; Bregman et al.2006), while the evidence for star formation is present at UV, optical, and IR wavelengths aswe presented in § II ]/H α , [S II ]/H α , and [O I ]/H α line ratios as a function of the velocitydispersion for nuclei and filaments in our sample. Point types/colors are described in Figure3. In the left panels, the regions occupied by LIRGs and ULIRGs (Monreal-Ibero et al. 2010)are shown in grey and yellow, respectively. In the right panels, we show two-line fits to thedata, with the Pearson R of each line in the upper left corner. Also shown in the upperleft corner is the F-test probability, P L , with which we can reject the null hypothesis of asingle-line fit in favor of a two-line fit. 29 –formation is both observationally motivated and straightforward to interpret. In order to testthis model, we turn to Eq. 1, which predicts the H α luminosity given an ICM cooling rate.We make the assumption that the cooling ICM surrounds the cool clouds and, thus, onlyhalf of the ionizing radiation is incident on the cool gas. We then compute the expected H α luminosity on a cluster-by-cluster basis, using X-ray-derived cooling rates from McDonaldet al. (2010) and McDonald et al. (2011a). Finally, we compare the predicted H α luminosityto the measured values from the aforementioned references and determine the fraction of H α luminosity in each cluster that is due to the condensing ICM. We further assume that thisfraction is constant throughout the cluster and remove the contribution to each emission lineflux from the cooling flow, assuming the predicted line ratios shown in Figure 14.Fig. 17.— Similar to figure 3, but with the expected contribution from the cooling flowremoved. By subtracting the expected flux based on the X-ray cooling properties of eachcluster, we find a better match between the optical emission line ratios in cool core clusterfilaments and typical H II regions, with the exception of the [O I ]/H α ratio.In Figure 17 we again show various optical line ratios for the nuclei and filaments in oursample of 9 clusters, but now with the predicted contribution from the cooling flow removed.We find that the [N II ]/H α and [S II ]/H α ratios are in better agreement with those for typicalH II regions in most systems, with the exception of two systems classified as shock-heated byMcDonald et al. (2011b), based on their H α /FUV flux ratios. However, the [N II ]/H α and[O I ]/H α ratios are still too high in most cases. Nevertheless, we feel that this compositemodel is a useful step forward since it introduces no additional free parameters and tries toaccount for a source of ionization which we know is present in cool core clusters. The opticalline fluxes due to the cooling ICM are tied directly to the X-ray-derived cooling rates andthe observed H α luminosities. 30 – Figure 7 shows that the line-emitting gas in cool core clusters generally has a coherentvelocity field. The combination of a bulk flow on the order of v r ∼
200 km s − and linewidths of a few times 100 km s − is reminiscent of the slow shocks combined with star-forming regions seen in large-scale winds (Rich et al. 2010) and in luminous infrared galaxies(Rich et al. 2011). This composite model, proposed by Farage et al. (2010) for the opticalfilaments in Centaurus A, produces slightly-elevated [N II ]/H α and [S II ]/H α , and highly-elevated [O I ]/H α , just as we observe in the cool cores of galaxy clusters. In Figure 18we show the predicted line ratios from a mix of star formation and radiative shocks fromAllen et al. (2008). We find that a modest fractional contribution from shocks ( ∼ § ∼ Z = Z (cid:12) . The gray background points are data forfilaments (stars) and nuclei (circles) in this sample. The upper panels show the measuredflux ratios, while the lower panels show flux ratios with the contribution from a self-ionizingcondensing ICM remove ( § I ]/H α ratios seen in the cool cores of galaxy clusters with a modest ( ∼ II ]/H α , and a weak or no correlation between the low ionization lineratios and velocity dispersion (star formation dominated; Abell 0478, Abell 0780, Abell 1795,Abell 2597, Sersic 159-03) and systems for which there is no detectable UV emission, high[N II ]/H α ratios, and a strong correlation between the low ionization line ratios and ve-locity dispersion (shock dominated; Abell 0496, Abell 1644, Abell 2052, NGC 4325). Thistable clearly demonstrates the differences between the star formation and shock dominatedsystems, in terms of their optical line ratios, optical line widths, UV flux and morphology,and nuclear line ratios. We caution, however, that the separation seen in Figure 16 is likelya selection effect and we would expect to see a continuum of systems with shock fractionsvarying from 0–100% in a larger, more complete sample.This scenario offers a straightforward explanation for, it seems, all of the observedproperties of optical filaments in cool core clusters. The combination of shocks and starformation account for the soft X-ray, UV and excess blue emission, mid-IR emission, slightlyelevated UV/H α ratios (McDonald et al. 2011b), small linewidths, and elevated [O I ]/H α ratios. Including contributions from the cooling ICM reduces the need for shocks by roughlya factor of two, but is not necessary to match the data. Assuming that the H II regionsare flowing along the filaments, they should be experiencing weak shocks due to interactionswith the relatively stationary ICM. The three most likely scenarios for the origin of the cool filaments in cluster cores areradial infall, buoyant outflow behind radio bubbles, and entrainment in radio jets. Thesemodels predict different kinematic signatures in the filaments which should be observable. Inthe classical picture of a rising bubble, cool material is entrained behind the bubble and risesto larger radius. Directly behind the bubble, the cool material should be moving away fromthe cluster center, while at small radius the gas which is trailing far behind the bubble coolsand falls back into the bottom of the potential well (e.g., Churazov et al. 2001; Reynoldset al. 2005). This results in a stretching of the filament, with the inner and outer portionsmoving in different directions. The line-of-sight velocities of such a filament should show achange of sign along the filament, as is seen in the core of Perseus A (Fabian et al. 2003;Hatch et al. 2006). While we do not observe a change in direction along any of the extendedfilaments in our sample, this may simply mean that we are observing the filaments earlyon, at which time the bulk of the gas will be outflowing, or at a late stage, when all of 32 –
Star Formation Dominated Shock DominatedSystems: A0478, A0780, A1795, A2597, A0496, A1644, A2052, N4325S159-03Optical < [N II]/H α > nuc = 0.99 ± < [N II]/H α > nuc = 2.70 ± < [S II]/H α > nuc = 0.89 ± < [S II]/H α > nuc = 1.59 ± < [O III]/H β > nuc = 0.25 ± < [O III]/H β > nuc = 0.28 ± < [N II]/H α > fil = 0.85 ± < [N II]/H α > fil = 1.76 ± < [S II]/H α > fil = 0.61 ± < [S II]/H α > fil = 0.82 ± α ∝ σ . [N II]/H α ∝ σ . [S II]/H α ∝ σ . [S II]/H α ∝ σ . UV Properties a : UV-H α flux/morphology correlation Weak UV emissionNuclear Properties b : Nuclear starburst LINERShock Fraction c : ∼ ∼ Table 2: Various properties of the two sub-samples which were defined based on how stronglythe low ionization line ratios were correlated with the velocity dispersion (Figure 16). a : For systems which have HST far-UV imaging from McDonald & Veilleux (2009) andMcDonald et al. (2011b). b : Based on optical line ratios in Figure 3. c : Fraction of H α luminosity produced by shocks, based on the composite star formation plusshocks models shown in Figure 18. the cool gas is falling back on the BCG. Indeed, the extended filaments in Abell 2052 andAbell 2597 are coincident with the outer rims of the X-ray cavities, but have nearly-constantline-of-sight velocites.A second scenario for the extended, cool gas is that it is being uplifted by radio jets.This was recently proposed as the origin of the extended warm gas in Sersic 150-03 (Werneret al. 2011), but was also investigated much earlier for a sample of radio-selected galaxies byBaum & Heckman (1989). In this scenario, cool gas which has settled in the cluster core isentrained within radio jets during an episode of feedback. Removed from the direct influenceof the AGN, this gas is allowed to cool further, resulting in the formation of young stars.The kinematics of the filaments in this case should be relatively smooth, with a constant 33 –velocity direction dictated by the radio jet. This is consistent with our observations for themost extended systems, as discussed in §
5. Summary
Utilizing a combination of narrow-band H α imaging and carefully-positioned long-slitspectroscopy we have obtained optical spectra of H α filaments in the cool cores of 9 galaxyclusters. These spectra provide a wealth of information about the warm, ionized gas incool core clusters, allowing us to provide the following new constraints on their origin andionization mechanisms: • We find evidence for only small amounts of reddening in the extended, line-emittingfilaments, with the majority of filaments having E ( B − V ) < . • We confirm that the optical line ratios [N II ] λ α , [O I ] λ α , [S II ] λλ α are slightly elevated above the expectation for star-forming regions. We findthat filaments identified as star-forming by their far-UV emission consistently have thelowest [N II ]/H α ratios. • For the complete sample there is a weak correlation between the low-ionization line 34 –ratios and line widths. This correlation is stronger for systems which appear to havea larger contribution from shocks than from star formation, based on their optical lineratios and far-UV/H α flux ratios. • The optical line ratios are able to rule out collisional ionization by cosmic rays, ther-mal conduction, and photoionization by both the ICM X-rays and AGN as strongcontributors to the ionization of the warm gas in both nuclei and filaments. • A composite model of a self-ionizing cooling plasma and star formation is able tobetter reproduce these line ratios than star formation alone, but still under-predictsthe low-ionization ratios in most systems. • The data are fully consistent with a combination of slow shocks and star formation, asseen in superwinds and merger remnants. This scenario explains the kinematics, lineratios, and multi-wavelength properties of the filaments and nuclei. We find typicalshock fractions of 0–40% in systems with signatures of star formation (i.e. far-UVemission), and 40–100% in systems with no observational evidence for star formation.Including self-ionization from the cooling plasma to this model reduces the necessarycontribution from shocks by roughly a factor of 2 while still matching the data. • The nuclei of Abell 0496, Abell 1644, and Abell 2052 are identified as LINERs. Theremaining 6 nuclei in the sample share similar properties to the star-forming filamentsand likely have very little contribution from an AGN. • In the central ( r <
10 kpc) regions of several cool core clusters, we see kinematicsignatures of rotating, H α structures, consistent with earlier work (e.g., Baum et al.1992). • The line-of-sight velocities of the most extended filaments peak at ∼
300 km s − . Wedo not observe a change in direction along any filaments. We provide evidence drawnfrom the literature for entrained gas in radio jets (Sersic 159-03, Abell 1795), coolgas uplifted by buoyant bubbles (Abell 2052, Abell 2597), and gravitational freefall(Abell 0496, Abell 1795). • The line profiles of the ionized filaments have widths on the order of FWHM ∼ − . This is consistent with velocity broadening due to slow shocks. In general,filaments that are thin and extended have narrower emission lines than those thatappear disturbed. • The line widths in the filaments decrease with distance from the BCG center. If thecool filaments have properties which reflect the ICM environment, this provides an 35 –estimate of the radial velocity dispersion profile in the ICM. These velocities widthsare consistent with those measured for the ICM using X-ray grating spectrometers(e.g., Sanders et al. 2011) and are much smaller than the stellar velocity dispersionsin the BCG.These new data, coupled with our previous work (McDonald et al. 2010, 2011a,b), sup-port a scenario where some fraction of the cooling ICM in the cluster core has been allowedto collapse and form stars. While in bulk motion (either infall or outflow), the cool gas isexperiencing shocks, producing the observed line widths of ∼
200 km s − , which are signif-icantly higher than those measured in typical H II regions. Further, the cooling ICM maycontribute a non-negligible fraction of self-ionization as it cools through EUV and UV transi-tions. In filaments which appear to be star-forming based on their FUV and IR emission, thecombination of shocks and self-ionization yield 0–40% of the H α luminosity in the filaments,with photoionization by young stars providing the additional 60–100%. Systems withoutany accompanying FUV emission appear to be fully consistent with ionization by radiativeshocks. While satisfactory, this scenario is necessarily incomplete. A complete modeling ofthe multiphase plasma using all available data from X-ray through radio, rather than focus-ing on a single energy regime, should allow for a more robust solution to this long-standingproblem, while combining the warm gas kinematics with deep X-ray and radio studies mayshed new light on the motions of this gas. Acknowledgements
Support for this work was provided to M.M. by NASA through SAO Award Number2834-MIT-SAO-4018, which is issued by the Chandra X-ray Observatory on behalf of NASAunder contract NAS8-03060, and to both MM and S.V. by NSF through contracts AST0606932 and 1009583, and by NASA through contract HST GO-1198001A. We thank R.Mushotzky and A. Edge for useful discussions. We also thank the technical staff at LasCampanas for their support during the ground-based observations.
REFERENCES
Abazajian, K. N., Adelman-McCarthy, J. K., Ag¨ueros, M. A., Allam, S. S., Allende Prieto,C., An, D., Anderson, K. S. J., Anderson, S. F., Annis, J., Bahcall, N. A., & et al.2009, ApJS, 182, 543 36 –Allen, M. G., Groves, B. A., Dopita, M. A., Sutherland, R. S., & Kewley, L. J. 2008, ApJS,178, 20Allen, S. W. 1995, MNRAS, 276, 947Allen, S. W., Edge, A. C., Fabian, A. C., Boehringer, H., Crawford, C. S., Ebeling, H.,Johnstone, R. M., Naylor, T., & Schwarz, R. A. 1992, MNRAS, 259, 67Baldwin, J. A., Phillips, M. M., & Terlevich, R. 1981, PASP, 93, 5Baum, S. A. 1992, PASP, 104, 848Baum, S. A., & Heckman, T. 1989, ApJ, 336, 702Baum, S. A., Heckman, T., & van Breugel, W. 1990, ApJS, 74, 389Baum, S. A., Heckman, T. M., & van Breugel, W. 1992, ApJ, 389, 208Bildfell, C., Hoekstra, H., Babul, A., & Mahdavi, A. 2008, MNRAS, 389, 1637Boehringer, H., & Fabian, A. C. 1989, MNRAS, 237, 1147Bregman, J. N., Fabian, A. C., Miller, E. D., & Irwin, J. A. 2006, ApJ, 642, 746Bregman, J. N., Miller, E. D., & Irwin, J. A. 2001, ApJ, 553, L125Canning, R. E. A., Fabian, A. C., Johnstone, R. M., Sanders, J. S., Crawford, C. S., Ferland,G. J., & Hatch, N. A. 2011, ArXiv e-printsCardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, ApJ, 345, 245Cardiel, N., Gorgas, J., & Aragon-Salamanca, A. 1995, MNRAS, 277, 502Churazov, E., Br¨uggen, M., Kaiser, C. R., B¨ohringer, H., & Forman, W. 2001, ApJ, 554,261Crawford, C. S., Allen, S. W., Ebeling, H., Edge, A. C., & Fabian, A. C. 1999, MNRAS,306, 857Donahue, M., de Messi`eres, G. E., O’Connell, R. W., Voit, G. M., Hoffer, A., McNamara,B. R., & Nulsen, P. E. J. 2011, ApJ, 732, 40Donahue, M., & Voit, G. M. 1991, ApJ, 381, 361Draine, B. T., & Salpeter, E. E. 1979, ApJ, 231, 77 37 –Dressler, A., Bigelow, B., Hare, T., Sutin, B., Thompson, I., Burley, G., Epps, H., Oemler,A., Bagish, A., Birk, C., Clardy, K., Gunnels, S., Kelson, D., Shectman, S., & Osip,D. 2011, PASP, 123, 288Edge, A. C. 2001, MNRAS, 328, 762Edwards, L. O. V., Hudson, M. J., Balogh, M. L., & Smith, R. J. 2007, MNRAS, 379, 100Edwards, L. O. V., Robert, C., Moll´a, M., & McGee, S. L. 2009, MNRAS, 396, 1953Egami, E., Misselt, K. A., Rieke, G. H., Wise, M. W., Neugebauer, G., Kneib, J., Le Floc’h,E., Smith, G. P., Blaylock, M., Dole, H., Frayer, D. T., Huang, J., Krause, O.,Papovich, C., P´erez-Gonz´alez, P. G., & Rigby, J. R. 2006, ApJ, 647, 922Fabian, A. C. 1994, ARA&A, 32, 277Fabian, A. C., Sanders, J. S., Allen, S. W., Crawford, C. S., Iwasawa, K., Johnstone, R. M.,Schmidt, R. W., & Taylor, G. B. 2003, MNRAS, 344, L43Farage, C. L., McGregor, P. J., Dopita, M. A., & Bicknell, G. V. 2010, ApJ, 724, 267Ferland, G. J., Fabian, A. C., Hatch, N. A., Johnstone, R. M., Porter, R. L., van Hoof,P. A. M., & Williams, R. J. R. 2009, MNRAS, 392, 1475Hansen, L., Jørgensen, H. E., Nørgaard-Nielsen, H. U., Pedersen, K., Goudfrooij, P., &Linden-Vørnle, M. J. D. 2000, A&A, 356, 83Hatch, N. A., Crawford, C. S., & Fabian, A. C. 2007, MNRAS, 380, 33Hatch, N. A., Crawford, C. S., Johnstone, R. M., & Fabian, A. C. 2006, MNRAS, 367, 433Heckman, T. M., Baum, S. A., van Breugel, W. J. M., & McCarthy, P. 1989, ApJ, 338, 48Hicks, A. K., Mushotzky, R., & Donahue, M. 2010, ApJ, 719, 1844Hu, E. M., Cowie, L. L., & Wang, Z. 1985, ApJS, 59, 447Jaffe, W., Bremer, M. N., & Baker, K. 2005, MNRAS, 360, 748Johnstone, R. M., Fabian, A. C., & Nulsen, P. E. J. 1987, MNRAS, 224, 75Kewley, L. J., Dopita, M. A., Sutherland, R. S., Heisler, C. A., & Trevena, J. 2001, ApJ,556, 121Kewley, L. J., Groves, B., Kauffmann, G., & Heckman, T. 2006, MNRAS, 372, 961 38 –Lim, J., Ao, Y., & Dinh-V-Trung. 2008, ApJ, 672, 252Mathews, W. G. 2009, ApJ, 695, L49McDonald, M., & Veilleux, S. 2009, ApJ, 703, L172McDonald, M., Veilleux, S., & Mushotzky, R. 2011a, ApJ, 731, 33McDonald, M., Veilleux, S., Rupke, D. S. N., & Mushotzky, R. 2010, ApJ, 721, 1262McDonald, M., Veilleux, S., Rupke, D. S. N., Mushotzky, R., & Reynolds, C. 2011b, ApJ,734, 95McNamara, B. R., & O’Connell, R. W. 1989, AJ, 98, 2018Melnick, J., Gopal-Krishna, & Terlevich, R. 1997, A&A, 318, 337Monreal-Ibero, A., Arribas, S., Colina, L., Rodr´ıguez-Zaur´ın, J., Alonso-Herrero, A., &Garc´ıa-Mar´ın, M. 2010, A&A, 517, A28+Murphy, J. D., Gebhardt, K., & Adams, J. J. 2011, ApJ, 729, 129Nipoti, C., & Binney, J. 2004, MNRAS, 349, 1509O’Dea, C. P., Baum, S. A., Privon, G., Noel-Storr, J., Quillen, A. C., Zufelt, N., Park, J.,Edge, A., Russell, H., Fabian, A. C., Donahue, M., Sarazin, C. L., McNamara, B.,Bregman, J. N., & Egami, E. 2008, ApJ, 681, 1035Oegerle, W. R., Cowie, L., Davidsen, A., Hu, E., Hutchings, J., Murphy, E., Sembach, K.,& Woodgate, B. 2001, ApJ, 560, 187Oke, J. B., Cohen, J. G., Carr, M., Cromer, J., Dingizian, A., Harris, F. H., Labrecque, S.,Lucinio, R., Schaal, W., Epps, H., & Miller, J. 1995, PASP, 107, 375Osterbrock, D. E. 1989, Astrophysics of gaseous nebulae and active galactic nuclei, ed.Osterbrock, D. E.Peterson, J. R., Kahn, S. M., Paerels, F. B. S., Kaastra, J. S., Tamura, T., Bleeker, J. A. M.,Ferrigno, C., & Jernigan, J. G. 2003, ApJ, 590, 207Quillen, A. C., Zufelt, N., Park, J., O’Dea, C. P., Baum, S. A., Privon, G., Noel-Storr, J.,Edge, A., Russell, H., Fabian, A., Donahue, M., Bregman, J. N., McNamara, B. R.,& Sarazin, C. L. 2008, ApJS, 176, 39Rafferty, D. A., McNamara, B. R., Nulsen, P. E. J., & Wise, M. W. 2006, ApJ, 652, 216 39 –Revaz, Y., Combes, F., & Salom´e, P. 2008, A&A, 477, L33Reynolds, C. S., McKernan, B., Fabian, A. C., Stone, J. M., & Vernaleo, J. C. 2005, MNRAS,357, 242Rich, J. A., Dopita, M. A., Kewley, L. J., & Rupke, D. S. N. 2010, ApJ, 721, 505Rich, J. A., Kewley, L. J., & Dopita, M. A. 2011, ApJ, 734, 87Sabra, B. M., Shields, J. C., & Filippenko, A. V. 2000, ApJ, 545, 157Salom´e, P., & Combes, F. 2003, A&A, 412, 657Sanders, J. S., Fabian, A. C., & Smith, R. K. 2011, MNRAS, 410, 1797Schlegel, D. J., Finkbeiner, D. P., & Davis, M. 1998, ApJ, 500, 525van Breugel, W., Heckman, T., & Miley, G. 1984, ApJ, 276, 79Veilleux, S., Kim, D.-C., & Sanders, D. B. 1999, ApJ, 522, 113Veilleux, S., Kim, D.-C., Sanders, D. B., Mazzarella, J. M., & Soifer, B. T. 1995, ApJS, 98,171Veilleux, S., & Osterbrock, D. E. 1987, ApJS, 63, 295Veilleux, S., Weiner, B. J., Rupke, D. S. N., McDonald, M., Birk, C., Bland-Hawthorn, J.,Dressler, A., Hare, T., Osip, D., Pietraszewski, C., & Vogel, S. N. 2010, AJ, 139, 145Voigt, L. M., & Fabian, A. C. 2004, MNRAS, 347, 1130Voit, G. M., & Donahue, M. 1990, ApJ, 360, L15—. 1997, ApJ, 486, 242Voit, G. M., Donahue, M., & Slavin, J. D. 1994, ApJS, 95, 87Werner, N., Sun, M., Bagchi, J., Allen, S. W., Taylor, G. B., Sirothia, S. K., Simionescu, A.,Million, E. T., Jacob, J., & Donahue, M. 2011, MNRAS, 415, 3369
This preprint was prepared with the AAS L A TEX macros v5.2.
40 – A pp e nd i x A W e p r o v i d e h e r e t h e r e s u l t s o f o u r s p ec tr o s c o p i c a n a l y s i s f o r e a c h s li t p o s i t i o n . F o r e a c h s p a t i a l e l e m e n t , w e p r o v i d e t h e r a d i u s f r o m t h e B C G nu c l e u s , t h e li n e - o f - s i g h t v e l o c i t y r e l a t i v e t o t h e B C G , t h e F W H M o f t h e H α e m i ss i o n li n e a ndflu x e s f o rt h e H α , H β ,[ O III ],[ O I ],[ N II ], a nd [ S II ]li n e s . T h e s e li n e s a r e un c o rr ec t e d f o r a n y r e dd e n i n g . Sp a t i a l b i n s w h i c hh a v e n o m e a s u r a b l e o p t i c a l e m i ss i o n a r e n o t s h o w n , f o r b r e v i t y . T a b l e A - :: K i n e m a t i c s a ndflu x e s ( un c o rr ec t e d f o rr e dd e n i n g ) d e r i v e d f r o m o p t i c a l s p ec t r a f o r e a c h s p a t i a l e l e m e n t a l o n g b o t h s li t s f o r e a c h ga l a xy c l u s t e r . N a m e S li t Sp a t i a l R a d i u s v r a d F W H M F H α F H β F [ O III ] F [ O I ] F [ N II ] F [ S II ] N o . B i n [ k p c ][ k m s − ][ − e r g s − ] A . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) C o n t i nu e d o nn e x t p ag e
41 – T a b l e A - c o n t i nu e d f r o m p r e v i o u s p a g e N a m e S li t Sp a t i a l R a d i u s v r a d F W H M F H α F H β F [ O III ] F [ O I ] F [ N II ] F [ S II ] N o . B i n [ k p c ][ k m s − ][ − e r g s − ] A . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) A . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) A . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) S . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) C o n t i nu e d o nn e x t p ag e
42 – T a b l e A - c o n t i nu e d f r o m p r e v i o u s p a g e N a m e S li t Sp a t i a l R a d i u s v r a d F W H M F H α F H β F [ O III ] F [ O I ] F [ N II ] F [ S II ] N o . B i n [ k p c ][ k m s − ][ − e r g s − ] . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) A . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) - . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) A . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) C o n t i nu e d o nn e x t p ag e
43 – T a b l e A - c o n t i nu e d f r o m p r e v i o u s p a g e N a m e S li t Sp a t i a l R a d i u s v r a d F W H M F H α F H β F [ O III ] F [ O I ] F [ N II ] F [ S II ] N o . B i n [ k p c ][ k m s − ][ − e r g s − ] . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) A . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) N . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . - ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( ) ( ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . ) . ( . )
44 –
Appendix B