The Relationship Between the Dense Neutral and Diffuse Ionized Gas in the Thick Disks of Two Edge-On Spiral Galaxies

2.9 MB·English

by  Katherine M. Rueff

Accepted for publicationin AJ
THE RELATIONSHIP BETWEEN THE DENSE NEUTRAL AND DIFFUSE IONIZED GAS IN THE THICK
DISKS OF TWO EDGE-ON SPIRAL GALAXIES 1
Katherine M. Rueff2, J. Christopher Howk2, Marissa Pitterle2, Alec S. Hirschauer2,3, Andrew J. Fox4, Blair
D. Savage5
Accepted for publication in AJ
ABSTRACT
Wepresenthigh-resolution,opticalimages(BVI+Hα)ofthemultiphaseinterstellarmedium(ISM)
3 inthethickdisksoftheedge-onspiralgalaxiesNGC4013andNGC4302. OurimagesfromtheHubble
1 Space Telescope (HST), Large Binocular Telescope, and WIYN 3.5-m reveal an extensive population
0 offilamentarydustabsorptionseentoz ∼2−2.5kpc. Manyofthesedustythickdiskstructureshave
2
characteristicsreminiscentofmolecularcloudsfoundintheMilkyWaydisk. OurHαimagesshowthe

n extraplanardiffuse ionized gas (DIG) in these galaxiesis dominated by a smooth, diffuse component.
a The strongly-filamentary morphologies of the dust absorption have no counterpart in the smoothly
J distributed Hα emission. We argue the thick disk DIG and dust-bearing filaments trace physically

2 distinct phases of the thick disk ISM, the latter tracing a dense, warm or cold neutral medium. The
dense, dusty matter in the thick disks of spiral galaxies is largely tracing matter ejected from the

] thin disk via energetic feedback from massive stars. The high densities of the gas may be a result of
A
converging gas flows. This dense material fuels some thick disk star formation, as evidenced by the
G presence of thick disk HII regions.
. Subject headings: dust, extinction – galaxies: individual (NGC 4013, NGC 4302) – galaxies: ISM –
h
galaxies: spiral – galaxies: structure – ISM: clouds
p
-
o
1. INTRODUCTION someadditionalmaterialmaybecontributedthroughin-
r
t fallingmatter;seeSancisietal. 2008). Thus,thestudyof
s The feedback of energy and matter from stars to the
a interstellarmedium(ISM)playsacentralroleintheevo- the statistical properties (e.g., Rossa & Dettmar 2003a,
[ 2003b; Howk & Savage 1999) and detailed physics (e.g.,
lution of galaxies. It impacts the physical/thermalstate

1 of the ISM, affecting the ability of the gas to form stars Haffner et al. 2009) of the thick disk ISM can help us
understandthe role that feedback plays in disk galaxies.
v and affects the galactic-scale star formation by ejecting
The energetic processes that expel gas from the thin
5 material from the thin disk or altogether from a galaxy
disk also act on the solid phase of the ISM, interstellar
2 (e.g., Dekel & Silk 1986). Most disk galaxies with suffi-
dust grains. This has been utilized to study the physi-
3 cientstarformationrates (orstarformationratesurface
cal structure as well as the phase structure of the thick
0 densities), show extended “thick disks” of warm inter-
disk ISM (Howk & Savage 1997,1999,2000;Alton et al.
. stellar matter extending ∼ 2 kpc away from the galaxy
1 2000,Thompsonetal.2004). While the energeticexpul-
mid-plane (Rossa & Dettmar 2003a, 2003b). The pres-
0 sion of matter from the disk can impact the dust-to-gas
ence of a thick disk ISM supported against gravity is
3
ratio through dust destruction, the fact that prevalent
a direct result of energetic feedback processes (although
1
extraplanardustisseenin∼50%ofnormalspiralgalax-
:
v ies (Rossa & Dettmar 2003a; Howk & Savage 1999) and
i 1BasedonObservationsobtainedwiththeNASA/ESAHub- eveningalacticsuperwindssuchasseeninM82(Kaneda
X bleSpaceTelescopeoperatedattheSpaceTelescopeScienceIn-
etal. 2010;Rousselet al. 2010;Engelbrachtet al. 2006;
stitute, whichisoperated bythe Association ofUniversities for
r ResearchinAstronomy,Inc.,underNASAcontractNAS5-26555. Hoopes et al. 2005; Scarrott et al. 1991) demonstrates
a Also, based on data acquired using the Large Binocular Tele- that dust can survive the ejection from the thin disk.
scope(LBT). TheLBTisaninternationalcollaborationamong
In fact, the detection of thick disk dust via the extinc-
institutions in the US, Italy, and Germany. LBT Corporation
partnersaretheUniversityofArizona,onbehalfoftheArizona tion it produces against background starlight in direct
universitysystem;InstitutoNazionaledoAstrofisica,Italy;LBT broadband images of edge-on galaxies is the most ob-
Beteiligungsgesellschaft,Germany,representingtheMaxPlanck servationally efficient method for detecting extraplanar
Society, theAstrophysicalInstituteofPostdam,andHeidelberg
interstellar matter (Howk & Savage 2000).
University; Ohio State University, and the Research Corpora-
tion, on behalfof theUniversityof NotreDame,the University The first detailed discussion of the implications of ex-
ofMinnesota,andtheUniversityofVirginia. Also,basedonob- traplanar dust was in the Howk & Savage (1997) (here-
servations obtained by the WIYN Observatory which is a joint after HS97) study of the nearby edge-on spiral galaxy
facilityoftheUniversityofWisconsin-Madison,YaleUniversity,
NGC 891, although the presence of high-z dust clouds
IndianaUniversity,andtheNationalOpticalAstronomyObser-
vatories. in this galaxy were noted earlier (Keppel et al. 1991;
2Department of Physics, University of Notre Dame, Notre Sandage 1961). HS97 showed that the thick disk of
Dame,IN46556. krueff@nd.edu NGC 891 was threaded with a network of dusty clouds
3DepartmentofAstronomy,IndianaUniversity,Bloomington,
and filaments observed to heights up to z ∼2 kpc. The
IN47405.
4Space Telescope Science Institute, 3700 San Martin Drive, most remarkable implication of this work was that the
BaltimoreMD21218. individual dust-bearing clouds have very large masses,
son5,DMepaadristomne,nWt oIf5A37s0t6ro.nomy, University of Wisconsin, Madi- often in excess of 105M⊙. Howk & Savage (1999) (here-
2 Rueff et al. 2012
afterHS99)subsequentlysurveyedagroupofninenearby olution. NGC 891 has one of the brightest known Hα-
edge-on spiral galaxies, demonstrating that high-z dust emitting DIG layers (e.g., Rand 1996, Miller & Veilleux
isfoundinmanygalaxiesandiscorrelatedwiththepres- 2003)andhas moreprominentfilaments within the DIG
ence of extraplanardiffuse ionizedgas(DIG) seen inHα than many galaxies (Heald et al. 2007). For galaxies
imaging. Subsequently,Rossa&Dettmar(2003a,2003b) with lower star formation rates, where the brightness of
showed that ∼ 40% of a much larger sample of spiral theDIGandtheprominenceoffilamentsdiminishes,are
galaxies exhibit extraplanar dust and verified the cor- there correspondingchangesin the extraplanardust? In
relation between the presence of extraplanar dust and this work we present high-resolutionimages of extrapla-
DIG emission. HS99 and Howk & Savage (2000; here- nardustandDIGintheedge-ongalaxiesNGC4013and
after HS00) argued that the gas traced by extraplanar NGC 4302. Our broadband images include observations
extinction features represented a much denser medium from HST that elucidate the smallest scale structures in
thanthatseeninHαemission,withdensitiesanorderof thedustythickdisk,whileourHαimagesareamongthe
magnitude or more higher than DIG material (see also highest spatial resolution acquired from the ground and
Keppeletal.1991). Theextraplanardustcloudsandfil- have good continuum subtraction. We use these images
aments,whicharenotspatiallycorrelatedwiththeDIG, to extend the detailed comparison of extraplanar dust
areonlyvisiblebecausetheyhavesignificantlyhigherex- and DIG properties to these two galaxies, which have
tinction–andlikelyhighercolumndensitiesandparticle fainter DIG emission and a smaller fraction of the emis-
densities – than their surroundings. HS00 hypothesized sion arising in filaments (i.e., a larger proportion com-
that the dust structures trace a cold, neutral medium ing from the smooth DIG component). In the end we
(CNM) in the thick disk basedon the largecolumn den- find no strongphysicalconnectionbetweenthe DIG and
sities, particle densities, and masses of the clouds. thestructurestracedbyhigh-contrastdustabsorptionin
This conclusion was motivated in part by a detailed these galaxies, similar to what is seen in NGC 891. The
comparison of the extraplanar dust filaments and DIG DIGanddust-ladencloudstracefundamentallydifferent
emission by HS00, which demonstrated that they were material in the thick disk.
tracing independent structures in the thick disk ISM of Our work is organized as follows. We discuss the ob-
NGC 891.6 Following the study of NGC 891,Thompson servationsanddatareductionin§2. In§3wediscussthe
et al. (2004; hereafter THS04), used the Hubble Space properties of the extraplanar dust and associated gas,
Telescope (HST) to investigate the extraplanar dust in while in §4 we describe the properties of the extraplanar
the edge-on galaxy, NGC 4217. These authors indicate Hα emission (DIG). In §5 we provide a direct compar-
that in addition to the smoother distribution of the Hα ison of the extraplanar dust morphologies and the DIG
emission morphology compared to the extraplanar dust detectedinHα. Wediscussin§6theorigins,nature,and
structures, they find little spatial correlation between propertiesofthe multiphaseISMinbothNGC 4013and
areas of enhanced Hα emission around regions of con- NGC 4302andsummarize the major results ofour work
centrated OB associations and high-z dust structures. in §7.
Rossa et al. (2004) similarly used images from (HST) to
show the ISM traced by extraplanar dust was distinct 2. OBSERVATIONS AND REDUCTIONS
in morphology from the DIG in NGC 891, while Rossa
The work in this paper is based on optical imaging of
& Dettmar (2003a, 2003b) argued for a similar arrange-
twogalaxies–NGC4013andNGC4302–whoseproper-
ment in their larger sample of galaxies. The dusty ex-
tiesarelistedinTable1. Weusebroadbandandnarrow-
traplanarfilamentsarephysicallydistinctfromtheDIG,
band imaging from several sources: 1) HST/Wide Field
even though the presence of these two tracers of the ex-
Planetary Camera 2 (WFPC2)) images of both galaxies
traplanar ISM is strongly correlated (Rossa & Dettmar
in the (WFPC2) equivalents of the BVI bands to study
2003b; HS99). The small scale structures in the HS00
extraplanardustabsorptionatthe highestresolution;2)
and subsequent Rossa et al. (2004) imaging, the high
ground-basedbroadbandimagesofNGC 4013andNGC
columndensities(N &1021 inmanycases),largecloud
H 4302takenwiththeWIYN3.5-mandLBT2×8.4-mtele-
masses, and high particle densities (n & 1−25 cm−3;
H scopes, respectively, to study extraplanar dust at larger
HS00,Howk2005)areallfeaturesingreatcontrasttothe
radialandverticaldistancesfromthegalaxycentersthan
otherobservedphasesofthe thickdiskISMandpointto
allowedby HST; and 3) narrow-bandHα imaging of the
adenser,coolermediumthanthe DIG.Thus,HS00con-
extraplanarDIGinthesegalaxies,takenwiththeWIYN
cluded that all of the thermal phases of the ISM found
3.5-m,to study the connection(or lackthereof)between
in the thin disks of spiral galaxies were also present in
the ionized gas and material traced by dust absorption
their thick disks.
in the thick disks of these galaxies. We discuss the pro-
Theonlygalaxyforwhichadetailedcomparisonofthe
cessing of these datasets in this section.
Hα (DIG) and extraplanar dust morphologies has been
undertaken is NGC 891. While the structures traced
by these two probes of the extraplanarISM were almost
completely distinct in NGC 891, the universality of the
2.1. HST/WFPC2 Observations and Processing
result is not settled. In particular, Hα images of the
DIG are often taken at relatively low spatial resolution, We obtained broadband imaging observations of
in part due to the faintness of the emitting gas, while NGC4013andNGC4302withtheWideFieldPlanetary
the dust structures are lost to confusion at similar res- Camera 2 (WFPC2) aboard HST. WFPC2, described
fully in Biretta et al. (2002), consists of four cameras.
6 Keppeletal. (1991)reachedthesameconclusiononthebasis The three Wide Field Cameras (WFCs) together cover
ofdensitiesestimatedfortheDIGandextraplanardustfeatures. an “L”-shaped region 150′′×150′′ with intrinsic spatial
3
Fig.1.—Threecolor(BVI)compositemosaicofNGC4013 createdfromtheHSTandWIYN imagingofthisgalaxy. Thefieldofview
ofthisimageis∼5′×5′ onaside. Northisup,eastistotheleft. Dustythickdiskcloudsareseenasdarkpatchesofextinctionoffofthe
galacticplane. TheHST/WFPC2imagesofthisgalaxycover thecentralandnortheastern sectionofthegalaxybutdonotincludemuch
ofthesouthwestern partofthe galaxy. Thebrightemissionintheplanejustnortheast ofthebulgeisaforegroundstar. Thereareslight
edgeartifactsduetothemergingoftheHSTandWIYNimages.
sampling of 0′.′1 per pixel (f/12.9). The Planetary Cam- The images acquired for each pointing were split
era (PC) covers a 34′′×34′′ square field with 0′.′046 per equallybetweentwoditheredsub-positions,withalinear
pixel (f/28.3). ditherstepof3.5WFCCDpixelsbetweenthepositions.
Our observations, summarized in Table 1, were ac- We usedthe Dither II packageandits drizzlealgorithms
quired as part of GO program 8242 (PI: Savage). We (see Fruchter et al. 1998 and Fruchter & Hook 1997) to
imaged the majority of the galaxy NGC 4302 in one combine the dithered images onto a finer grid, following
pointing, while we used two separate pointings to cover THS04. The resulting cosmic ray-cleaned images have
a significant fraction of NGC 4013. For NGC 4302 and an angular sampling of 0′.′05 per pixel.
one of the NGC 4013 pointings, we acquired eight ex- Figures 1 and 2 show composite color images of
posures in the F450W filter and four exposures in each NGC4013andNGC4302,respectively,createdfromthe
of the F555W and F814W filters. We obtained four WFPC2 BVI images. The WFPC2 images have been
images in each of the three filters for the secondary combined with ground-based observations in these im-
NGC 4013 pointing. The F450W, F555W, and F814W ages. The ground-based images are more sensitive to
filters roughly correspond to the Johnson-Cousins B, V, faint, extended structures (i.e., the faint thick disk stel-
and I bands7. lar emission) and probe a larger field of view. For NGC
4013, the ground-based data are from the WIYN 3.5-
7 We chose the F450W filter as the B-band equivalent because mtelescopeandarediscussedin§2.2. Theground-based
it has a higher over-all throughput than the F439W filter; the
transformation to the standard Johnson-Cousins system is well- behaved(seeHoltzmanetal.1995).
4 Rueff et al. 2012
Fig.2.—Threecolor (BVI) compositemosaicimageofNGC 4302 created fromtheHSTandLBT imagingof thisgalaxy. The fieldof
viewofthisimageis∼8′×5′ onaside. Northistotheright,eastisup. Dustythickdiskcloudsareseenasdarkpatchesofextinctionoff
of the galactic plane. The HST/WFPC2 images of this galaxy cover the central region out to R ∼6 kpc. NGC 4302 shows no signature
of interaction withthe neighboring galaxy, NGC 4298, seeninthe bottom of the composite image. Thereareslightedge artifacts due to
themergingoftheHSTandLBTimages.
observationsofNGC4302wereobtainedwiththe2×8.4- that described in HS00. We note that because the field
mLargeBinocularTelescope(LBT)andaredescribedin of view of the imager is relatively small compared with
§2.3. the galaxies themselves, the sky backgroundsubtraction
is somewhatuncertain. None of our conclusions are sen-
2.2. WIYN Observations and Processing sitivetothisproblem;however,itprecludesusfrom,e.g.,
properly deriving scale heights of the diffuse ionized gas
To study the distribution of the diffuse ionized gas in
in these galaxies.
thesegalaxies,weacquirednarrow-bandHαandGunnr
We derived the astrometric plate solution for our im-
imagesofNGC 4013andNGC 4302atthe WIYN 3.5-m
ages using a grid of stars whose coordinates were mea-
telescopeatKittPeakNationalObservatory. Thesedata
sured in the Digitized Sky Survey for the NGC 4302
were collected on 2004April 18 and 19 (UT) under non-
observations. These agree well with the HST-provided
photometric conditions using the WIYN Tip-Tilt Mod-
plate solution for the WFPC2 data. The rms error in
ule (WTTM),althoughwithoutthe tip-tiltfunctionality
using the plate solution to derive coordinates is always
enabled. A log summarizing the WIYN observations is
significantly less than 1′.′0. For the NGC 4013 observa-
given in Table 1, including the total exposure time in
tions, this treatment of the data left a systematic offset
each filter and the seeing-limited resolution in the final
betweentheWIYNandHSTdata. DuetoalackofDSS-
images for each galaxy.
TheWTTM8 is anopticalre-imagingsystemworking detected stars in the small WFPC2 field, we used stars
andgalaxiesincommonbetweentheWFPC2andWIYN
at f/7.43. This system feeds a 2048×4096 EEV CCD
with 13.5 µm pixels. Each pixel subtends 0′.′1125 on the images to bring the WIYN data onto the same reference
sky giving a 3.′8×4.′7 field of view that is vignetted at frame as the HST data. The BVI imaging of NGC 4013
was obtained with the previous generation WIYN Im-
the corners. The data acquisition, calibration, and re-
ager, the S2KB Camera. The properties and reductions
duction of the images were done in a manner similar to
ofthese dataareasthosediscussedinHS00. Thesedata
8 The WTTM instument described here: are also summarized in Table 1.
http://www.noao.edu/wiyn/WTTM manual.html. The principal purpose of our ground-basedWIYN ob-
5
servations is to study the distribution of Hα emission Sky Survey (SDSS-DR7; Abazajian et al. 2009). This is
from the DIG in these galaxies. The WIYN W15 filter a necessary step due to the distortions over the full field
wasusedasanon-bandHαfilter. Thisfilter,centeredat given the f/1.14 focal ratio. The absolute astrometric
6569˚A with a FWHM of ∼73˚A, contains emission from calibration has an rms dispersion compared with SDSS
the Hα and nearby [NII]9 lines plus a stellar continuum of ∼ 0′.′2 across the entire field of view. The individual
contribution. We used the Gunn r observations to es- images were subsequently resampled and coadded using
timate the stellar continuum and subtract it from the SWARP.
narrow-band images, scaling the rband images so that
3. THE DUSTY THICK DISKS OF NGC 4013 AND
faint Milky Way foreground stars were completely re-
NGC 4302
movedfromthe Hαimagesaftercontinuumsubtraction.
Non-saturatedforegroundstarsandbackgroundgalaxies Visual inspection of Figures 1 and 2 reveals the pres-
arewellsubtractedinourfinallineimages. Wefoundno ence of strong, patchy extinction in the thick disks of
evidence for a need to adopt a different scale factor for these two galaxies. Filaments and clouds of absorbing
the bulge of the galaxies compared with the outer disks material are seen to nearly the radial edge of the galax-
ofthe galaxies(e.g.,Rand1996). TheGunnr filterused ies’opticaldisksandtoz-heightsapproachingthelargest
for the continuum estimate encompasses the Hα line it- distances to which stellar light can be seen in the rel-
self and may lead to slight over-subtraction of the line atively shallow WFPC2 images, although the ground-
emission. This has little impact on the currentpaper, in basedLBTandWIYN imagesrevealthe dustextinction
whichwelimitourselvestomorphologicalconsiderations. extendstolesserheightsthanthestellarthickdisk. This
extraplanarextinction is a direct tracerof the thick disk
2.3. LBT Observations and Processing ISM in these galaxies (e.g., Howk 2005). In order to be
seen,thesedustyregionsmusthavehighercolumndensi-
We obtained broadband UBVI images of NGC 4302
ties than their surroundings with these filaments having
withtheLargeBinocularCameras(LBCs)onthe2×8.4-
excess extinction compared with nearby sight lines, pro-
m LBT under non-photometric conditions on 2008 May
vidingsignificantcontrastbetweenthesecloudsandtheir
03. A log summarizing the LBT observations is given in
local environment in our images. Smoothly distributed
Table 1 including the total exposure time in each filter
dustdoesnotproducesuchcontrast;thusadiffuselydis-
and the seeing-limited resolution in the final images for
tributed component of dust may exist undetected in the
NGC 4302. The LBT uses two co-pointed 8.4-m mir-
thick disks of these galaxies. That the column densities
rors to feed a pair of nearly-identical LBCs. The LBCs
oftheobserveddust-bearingcloudsarehigherthantheir
aredescribedinGiallongoetal.(2008). Thetwocameras
surroundings almost certainly implies the dust features
operatesimultaneouslyatprimefocus,oneoptimizedfor
seen in our images have significantly higher volume den-
bluewavelengthsandtheotherforred(bothcanprovide
sities than their surroundings, as well.
V-band imaging, the approximate cross-overin sensitiv-
ity). Each camera provides a ∼ 23′ ×23′ field of view Figure 3 shows the HST and WIYN V-band images of
using a four CCD mosaic with 0′.′23 pixels. Inter-chip NGC 4013. The top panel shows the HST V-band data
and the middle panelshows the unsharpmaskedversion
spacingsare filled in by dithering the telescope oversev-
of the top panel. The display of the galaxy in the top
eralexposures. AlogofourobservationsisgiveninTable
panelisnotabletoshowtheabsorbingcloudsagainstthe
1. WeusedtheLBC-BluecameratoimageNGC4302in
bulgeandfurther outin a singledisplay,makingthe un-
the UandB filters,while theLBC-Redcameraprovided
sharp masking necessary. The unsharp masked versions
coverage of in the V and I filters. We use twilight sky
ofourimagesremovethe large-scalegradientsinthe un-
exposures to perform the flat field corrections.
derlyingstellarlight,revealingstructureonscalessmaller
Reductionofthe LBCimagesrequirescombininga se-
thanthe smoothingkernel(FWHM.350pc) (see HS00
ries of dithered images, taking into account the distor-
andTHS04fordetaileddiscussionsofthisapproach). We
tions that occur over the very wide field of view. We
only use these images for display and have taken care
largely follow the techniques described in Sand et al.
to assess and minimize masking-related artifacts. Both
(2009)using modified versionsof scripts from B. Weiner
HST images include insets showing the central bulge re-
et al. (in preparation). The data are processed in two
gionwith a stretchthat showsthe dust absorptionmore
steps. The firstuses standardreductionsfromthe IRAF
clearly. The bottom panel shows the unsharp masked
package mscred to trim and remove the overscan re-
display ofour final WIYN data. The WIYN image gives
gion, subtract bias frames, and apply flat fields (derived
almostfullradialcoverageofthegalaxy,revealingextra-
from twilight flats). Subsequently, we used the SCAMP
planar dust structures extending along nearly the entire
(Bertin 2002) and SWARP (Bertin et al. 2002) soft-
radiallengthofthe disk ofNGC 4013. Extraplanardust
ware packages10 to produce the final images. SCAMP
clouds can be identified against the background stellar
was used to derive the astrometric solution for each of
light z ∼2.0 kpc from the midplane. Those dust clouds
the chips in each image. The final solution is based on
at the largest heights show less prominent substructure
cross-matching sources identified in the LBT images us-
andgenerallyhavelowera values,buttheyarestillwell
ing SExtractor (Bertin & Arnouts 1996) with an astro- V
detected in our WIYN images. We see no evidence for
metric catalog derived from DR7 of the Sloan Digital
disparities in the amount of extraplanardust comparing
opposite sides of the plane.
9 While the images contain both Hα and [NII] emission, for
Similarly, Figure 4 shows the HST and LBT V-band
brevity’s sake we willhereafter referto our images as Hαimages.
The reader should be aware that the bandpass also contains the images for NGC 4302. The top panel shows the HST
[NII]transitions. V-band data and the middle panel shows the unsharp
10 http://www.astromatic.net/ masked versionof the top panel. The HST images cover
6 Rueff et al. 2012
Fig. 3.—HST&WIYNV-bandimagesofNGC4013. ThetoppanelshowstheHSTV-bandimage;themiddlepanelshowstheunsharp
masked version of this image. The insets in the top two panels are stretched to show the bulge region in greater detail. The area of the
insets is indicated by dashed lines. The bottom panel shows the unsharp masked version of the WIYN V-band image. The dashed box
inthe WIYN V-band indicates the extent of the HSTimage shown in the upper panels. Allimage displays areinverted so that areas of
dustextinction appearlighterthantheirsurroundings. TheHSTimagesshowdense, narrow,filamentary-likeextraplanarduststructures
athighresolution. TheWIYN imageshowthe extensive denseduststructures andcomplexes extending toheights z∼2kpcbeyondthe
midplane.
∼6 kpc projected radial distance from the center on ei- kpc from the midplane, detected in the LBT images.
ther side, while the LBT images show ∼ 13 kpc radial The FHWM of the Gaussian used to produce the un-
span of the optical disk and more. The dashed box in sharpmasked images for NGC 4302were the same as in
the LBT V-band indicates the extend of the HST image NGC 4013.
shown in the upper panels. Extraplanardust clouds can
be identified againstthe backgroundstellar light z ∼2.5
7
Fig. 4.—HST&LBTV-bandimagesofNGC4302. Thetoppanel showstheHSTV-bandimage;themiddlepanelshowstheunsharp
maskedversionofthisimage. ThebottompanelshowstheunsharpmaskedversionoftheLBTV-bandimage. ThedashedboxintheLBT
V-bandindicatestheextendoftheHSTimageshownintheupperpanels. Allimagedisplaysareinvertedsothatareasofdustextinction
appearlighterthantheirsurroundings. TheHSTimagesshowalargenumberofextraplanarduststructuresathighresolution. TheLBT
imageshows extraplanar dust-bearingclouds inthe thickdiskalongnearlythe entire radiallength ofthe galaxy andis moresensitive to
smallextinction clouds.
3.1. Quantifying Dust Extinction tinction, a , for a given wave band λ as
λ
UsingtheHSTV-banddataforourmeasurements(top
a =−2.5log(S /S ); (1)
panelsofFigure3and4)andfollowingHS00andTHS04, λ dc,λ bg,λ
we provide spatially-averaged estimates of the apparent
hereS is the averagesurfacebrightnessmeasuredto-
opacities of a small number of extraplanar dust struc- dc,λ
wardadustcloud(dc),andS isthesurfacebrightness
tures. From these we derive the “apparent” extinction, bg,λ
ofthelocalbackground(bg). Definedinthisway,theap-
column densities, particle density, and masses of these
parent extinction is a lower limit to the true extinction,
dust-bearing structures. We define the “apparent” ex-
A , because S includes both extincted starlightfrom
λ dc,λ
8 Rueff et al. 2012
behindaswellasunextinctedlightemittedinfrontofthe sumption about the gas-to-dustrelationshipin the thick
feature. One can, in theory, use a for multiple wave- disk. Following earlier works, we assume the clouds fol-
λ
lengths with an assumed extinction curve to estimate low dust to gas relationships appropriate for the disk
the fraction of light originating in front of the dust fea- of the Milky Way (HS97, HS99, HS00). This is a nec-
ture(e.g.,Gallagher&Hunter 1981,Knapenetal.1991, essary step for estimating the physical conditions of
HS97). In this case, the fraction of S arising in front the dust features from the measured apparent extinc-
dc,λ
of the clouds is x and the true extinction is a function tions, one whose consequencesare discussed at length in
of a and x. However, HS00 found that this approach HS00. We adopt the average relationship between to-
V
does not alwayslead to self-consistent results, likely due tal hydrogen and the color excess E(B −V) of diffuse
to the effects of spatial averaging of probing differing clouds in the Galactic disk as estimated by Bohlin et
depthsintocloudsatdifferentwavebandsandofscatter- al. (1978): N(HI+H )/E(B −V) = 5.8×1021 atoms
2
ing. HS97 (who used a narrower range of wavelengths) cm−2 mag−1. Torelatethistotheextinction,weassume
providedevidenceforasignificantamountofstellarlight R ≡A /E(B−V)≈3.1, roughlythe meanfor diffuse
V V
arising in front of the dust clouds in NGC 891. clouds in the Galactic disk (e.g., Cardelli et al. 1989).
HerewefollowHS00andTHS04inassumingAV &aV Thisgivesthe totalhydrogen(HI+H2)columndensity
(equivalent to assuming x≥0). This has the advantage from the apparent extinction:
that our calculations of the physical properties of the
structuresseeninourimageswillprovidelowerlimits to N (cm−2)>1.9×1021a . (2)
H V
the true properties. HS00 discuss the degree to which
AV departs from this assumption for several values of The inequality arises since aV < AV. Since both
x. For example, for x = 0.25, the V-band extinction NGC 4013 and NGC 4302 are relatively massive spiral
A ≈ 1.5a . We do not report the extinction in the galaxies,theassumptionofaMilkyWaygas-to-dustratio
V v
other bands. will probably yield a reliable estimate of the associated
For each dust feature we measure the ratio of sur- gascolumndensity. However,ifsignificantdust destruc-
facebrightnessesS /S usingintensitydistributions tion has occurred in any particular feature, the derived
dc,λ bg,λ
measured vertically or horizontally, with respect to the NH will be an underestimate, while if the dust has been
disk, averaging the results for a given cloud. We es- separatedfromthe gasthenNH willbe anoverestimate.
timated the background surface brightness appropriate Wefavorscenariosinwhichlittledustdestructionoccurs
for each feature by fitting a spline to the intensity dis- in moving the dust from the thin to thick disk or where
tribution of the star light in each cut. This technique somedestructionhas occurredinwhichthe dustis sepa-
is different than that used in our previous works where rated from the gas. We thus quote the column densities
we were more concerned with the impact of the smear- as lower limits.
ing due to the seeing. The current measurements will We give lower limits to the masses of for the features
averagetheaV valuesovermoreofthecloudsthanthose in Table 1, calculated as M ∼ µmHNHAdc, where Adc
earlier measurements, leading to somewhat smaller ex- is the projected area of the dusty clouds. We assume
tinction values. µ=1.37tocorrectforthecontributiontothemassfrom
helium and the heavy elements. The definition of the
3.2. Physical Properties of Individual Dust Structures cloudareasisacriticalandhighly-subjectivecomponent
Table 1 summarizes the physical properties of several of the mass estimates. In the present work we have fa-
cloud complexes in the thick disks of the two target vored small clouds and have been conservative in our
galaxies from the HST V-band data. Figures 5 and 6 derivation of the projected areas. The implied masses
identify the clouds discussed in Table 1, they are dis- should be treated as lower limits.
played in the unsharp-masked WFPC2 images to better In addition to the columns and masses, we also in-
showthesmall-scalefeaturesoftheselectedclouds. Sev- cludeaverycrudedensityestimateinTable1. Thisesti-
eral measured quantities are given, including their posi- mateassumesthe cloudsweseearelargelycylindrically-
tions, physical dimensions, height above the midplane, symmetric such that nH ∼ NH/∆x, where ∆x is the
and a values. Estimates of the column densities, den- width of the minor axis of the clouds. The only support
V
sities, and masses associated with the clouds are also for this assumption is the general prevalence of filamen-
given, as outlined below. In this work we have favored tary absorbing structures seen in our images, although
small,relativelyisolatedstructures inorderto study the this impression could be affected by the general confu-
densitiesimpliedbythesmall-scalestructureobservedin sioncausedbyoverlappingcloudsalongthe lineofsight.
the WFPC2 imaging. The choice to work with isolated In general the morphologies of the absorbing clouds are
cloudstendstofavorlowerextinctioncloudsatrelatively complex, but we move forward using this simplistic as-
large z. The smaller sizes (to emphasize high densities) sumption for estimating the densities.
tend to favor lower masses. Keeping in mind the uncertainties discussed above,
The values of a in Table 1 are in the range 0.1–0.3 the properties of the clouds summarized in Table 1 are
V
mag. Thesevaluesrepresentspatialregionsoftheclouds nonetheless intruiging. The features all have estimated
lower than the values appropriate for the cores of the columns NH &2×1020 cm−2, simply as a matter of se-
structures. Given the values of aV are also lower than lection (it is difficult to identify clouds with av < 0.1).
thetrueextinction,realisticvaluesofaV formuchofthe We infer cloud masses &104M⊙ for all but the smallest
clouds are likely significantly larger than those values twocloudsinTable1,andthedensitiesimpliedforthose
given in Table 1. cloudsincludedinthetableareofordern ∼1−4cm−3.
H
Transformingtheapparentextinctionsdiscussedabove Additionally, we examined some of the darkest narrow
into physical properties of the clouds requires an as- filaments in the WFPC2 images located in regions too
9
Fig.5.— HST unsharp masked V-band image of NGC 4013 with select dense extraplanar clouds outlined and labeled. The properties
of these clouds are summarized in Table 1. These particular absorbing structures were chosen based on their location, apparent opacity,
and sizeinan effort to emphasize the physical properties of smaller, dense structures that the high-resolutionHST images reveal. These
extraplanarcloudsarelocatedatheightsz∼430−930pc,haveestimatedcolumndensitiesandmassesNH>2×1020cm−2,andM &104
M⊙. Substructure inthese clouds is seen to the limitof our resolution. The orientation of this image is as in Figure3. Figure8 gives a
close-upimageofthegrayoutlinedclumptotheeastbelowthebulge,cloud115834.4+435651.9.
Fig.6.—HSTunsharpmaskedV-bandimageofNGC4302withselectdenseextraplanarcloudsoutlined. Theorientationofthisimageis
asinFigure2. ThepropertiesofthesecloudsaresummarizedinTable1. Theseparticularabsorbingstructureswerechosenbasedontheir
location, apparent opacity, and sizeinaneffort to emphasize the physical properties of smaller,dense structures that the high-resolution
HSTimages reveal. These extraplanar clouds, located at heights z >590 pc, have estimated columndensities and masses NH >2×1020
cm−2,andM >104 M⊙,respectively. Substructureinthesecloudsisseentothelimitofourresolution.
confused to show well in Figures 5 and 6, so they are smaller-scale filaments. The masses given in Table 1 are
not included in Table 1. Several of these clouds have smaller than have been reported in earlier works (HS99
column densities approaching or exceeding 1021 cm−2, HS00,THS04),butthatislargelyafunctionofournewly
giving mean densities n & 8−10 cm−3, which is ex- adoptedmeasurementtechniqueandthechoiceswehave
H
traordinary for thick disk clouds. The column densi- madeforemphasizingthe smallersubcomponentsrather
ties and densities measured for the thick disk clouds in than the largest complexes. Complexes of absorption
NGC 4013 and NGC 4302 are similar to those seen in with implied total masses well in excess of 105M⊙ cer-
NGC 891 and NGC 4217 (HS97, HS00, THS04, Howk tainly exist in these galaxies. Overall,we do not see any
2005). The derived densities likely depend on the reso- clear evidence that the dust-bearing thick disk clouds in
lution with which one observes the structures, as higher thesegalaxiesaresignificantlydifferentincharacterthan
resolution images not only smear less background light those seen in NGC 891 (HS00) or NGC 4217 (THS04).
into the structure but also allow the identification of
10 Rueff et al. 2012
Fig.7.—(a)AsectionofthebulgeofNGC4013fromtheHSTunsharpmaskedV-bandimagesshowingthelargedustcomplexdiscussed
in the text. (b) Close-up view of the large complex, which extends ∼900 pc by 200 pc. This complex has a mean extinction aV =0.04
and total mass & 105M⊙. (c) An enlarged view of two of the cometary-like clouds associated with the large complex. These are clouds
115832.2+435644.6and115832.0+435644.1inTable1. Cloud115832.2+435644.6hasanotablecore-halotypestructure,asemphasized
intherelativecontours shownonthisdisplay. ThecoreofthecloudhasaV =0.36andimplieddensitynH≥7cm−3. Theorientationof
theseimagesisasinFigure3.
3.3. Apparent morphologies of dust-bearing gas clouds ies. Ursa Major is considereda loose groupwith no cen-
tral concentration and no sign of strong HI deficiencies,
Ourimagesallowustoidentifylarge-scaledust-bearing
suggestingthisgrouplacksastrongintraclustermedium
cloudsinthethickdisksofthesegalaxiesandstudytheir
(Angiras et al. 2007). As a member of the Ursa Major
internal structure. Our high angular resolution HST
group,NGC4013isrelativelyisolated. Thusthebending
images reveal substructure within the cloud complexes,
of filaments is not likely the resultof movement through
down to the limit of our resolution, but only for the in-
an intracluster medium in this case (Mart´ınez-Delgado
ner sections of both galaxies. Our LBT and WIYN data
et al. 2009). We find no evidence for obvious supershell
allowustoidentifydustcloudsatlargerradialandverti-
structuresinNGC4013,althoughidentifyingsuchstruc-
caldistances due to their greaterfield ofview and depth
tureswithinthe complexdistributionofcloudsmightbe
comparedwiththeHSTimages. Similartocloudsinthe
difficult.
MilkyWay,inwhichstructureisseenonmanyscales,the
We see a set of very large-scale complexes of dust
thick disks of NGC 4013 and NGC 4302 contain narrow
above and below the bulge of the galaxy in Figure 3.
filamentarystructures,largecomplexesofclouds(specif-
These large dust complexes, traced out to z ∼ 1.8 kpc
ically in NGC 4302), as well as small, very dense clouds
on either side of the bulge, show significant small-scale
seen to our resolution limit.
structure within them, including dense clouds similar to
3.3.1. NGC 4013 those cometary shaped clouds identified in HS00. Fig-
The extraplanar dust-bearing clouds in NGC 4013are ure 7 shows one of these complexes to the southeast of
generally highly structured filaments, although we see a the bulge. This large complex (∼ 900 pc ×200 pc) ex-
broadrangeofmorphologiesinboththeHSTandWIYN tends from within ∼ 500 pc of the plane, bending to-
imagesinFigure3. The typicaldustfilamentsare∼100 ward the east with increasing height above the plane.
pc by a few ×100 pc in size (see Table 1 and Figure It shows several cometary-like substructures at heights
3). We note that the dust filaments are unlikely to be z ∼600−850 pc. Figure 7(b) provides a close-up image
associated with a flare in the outer galaxy (see detailed ofthis largecomplex,whichhasameanextinctionvalue
discussion in HS99, HS00) or as a result of the warp in of aV = 0.04 and total mass of order & 105M⊙. Figure
NGC 4013, which is perpendicular to our line of sight 7(c) shows two of the the cometary-like clouds, clouds
(Bottema 1996) and therefore would not give rise to the 115832.2+435644.6 and 115832.0+435644.1 from Ta-
symmetrical distribution of clouds seen in our images. ble 1. Contours are included to guide the eye (but are
Thereisastrongpreferenceforthehighest-zabsorbing basedon the unsharp maskedimage and are not quanti-
structures (z ≥ 2 kpc) in NGC 4013 to appear bent to- tatively useful). Cloud 115832.2+435644.6 has a head-
ward the northeast, especially on the eastern side of the tail type structure; the head of the cloud has aV =0.36
galaxy(see bottom panel in Figure 3). Suchan arrange- and implied density nH ≥ 7 cm−3. The association of
mentcouldoccurifthereisarotationallagthatincreases these cometary structures with the larger complex sug-
with height (e.g. Heald et al. 2007). This is consistent geststhecomplexisinteractingwithanambientmedium
with the sense of rotation of NGC 4013,with the south- asitfallstowardthediskthroughacoronaoroutflowing
westsiderecedingfromus(Bottema1996). NGC4013is wind from the disk.
a member of the spiral-rich Ursa Major group of galax- Allofthehead-tailcloudsweseeinNGC4013seemto