Radial structure of galactic stellar disks

a r

X

i

v

:a

s t

r o

-

p

h

/

2

1

10

v

1

2

J

a

n

2

2

Disks of Galaxies :Kinematics,Dynamics and Perturbations ASP Conference Series,Vol.iii,2002E.Athanassoula &A.Bosma,eds.Radial structure of galactic stellar disks M.Pohlen,R.-J.Dettmar,R.L¨u tticke,&G.Aronica Astronomisches Institut,Ruhr-Universit¨a t Bochum Universit¨a tsstr.150,D-44780Bochum,Germany Abstract.We present the results from our deep optical imaging survey μV lim ≈26?27mag /2′′of a morphologically selected sample of 72edge-on disk galaxies.The question of the global structure of galactic stellar disks,especially the radial surface brightness pro?le at large galactocen-tric distances,is addressed.We ?nd that typical radial pro?les are better described by a two-slope exponential pro?le —characterised by an in-ner and outer scalelength separated at a break radius—rather than a sharply-truncated exponential model.Results are given for three face-on equivalents,serving as the crucial test to assure the ?ndings for the edge-on sample without possible geometrical line-of-sight e?ects.1.Introduction Van der Kruit (1979)initially found that the outer parts of disks of spiral galax-ies do not retain their exponential light distribution to the observed faint levels,but rather show sharp edges.For three nearby edge-on galaxies (NGC 4244,NGC 4565,NGC 5907)he derived that the typical radial scalelength h steepens from 5kpc to about 1.6kpc at the edge of the disk.The existence of these trun-cations,which are already visible in contour maps of edge-on and even of some face-on galaxies,is now well accepted (Pohlen 2001),but no unique physical in-terpretation is given to describe this observational phenomenon.The proposed explanations span a rather wide range of possibilities.Van der Kruit (1987)de-

duced a connection to the galaxy formation process describing the truncations as remnants from the early collapse.Ferguson &Clark (2001),for example,pro-posed an evolutionary scenario represented by the viscous disk evolution mod-els.And Kennicutt (1989)suggested a —probably less striking—star-formation threshold.Up to now the applied characteristic parameter for comparing the observational results of di?erent studies is the distance independent ratio of a truncation radius R t to a measured radial scalelength h .Van der Kruit &Searle (1982)found for their sample of seven galaxies a value of R

t /h =4.2±0.6,

whereas Pohlen,Dettmar,&L¨u tticke (2000a)derived a signi?cantly smaller one of R t /h =2.9±0.7for their CCD survey of 30galaxies.

2.Our new edge-on sample

To explain these di?erent R t /h values,we have improved the rather inhomo-geneous sample of Pohlen et al.(2000b).The resulting morphologically se-

1

2M.Pohlen,R.-J.Dettmar,R.L¨u tticke,&G.Aronica

Figure1.Radial surface brightness pro?les(solid lines)of NGC522

(left panel)together with the best?tting sharply-truncated model

(dashed line)and the best?tting two-slope model(right panel).

lected sample contains72galaxies of high data quality.They are selected to be edge-on,undisturbed,and similar to’well-behaved’disk-prototypical cases such as NGC4565and IC2531.Thereby we want to assure that we are able to consistently?t our simple one-component disk model.We have chosen S0-Sd galaxies(mainly Sb-Sc)and have obtained deep optical imaging in at least one?lter,reaching a limiting surface brightness ofμ1σlim≈26?27V-mag/2′′or μ1σlim≈25?26R-mag/2′′.

We have?tted three di?erent kind of models to the data(cf.Pohlen2001):A sharply-truncated exponential model characterised by the cut-o?radius R co and the corresponding scalelength h co.The cut-o?radius is derived at the position where the pro?les bend vertically into the noise.An in?nite-exponential model, which is only characterised by a scalelength h∞and realised by?xing the cut-o?radius to ten times the sharply-truncated scalelength,to address the problem of the a priori assumed sharp truncations.And?nally a two-slope or smoothly-truncated model,characterised by a break radius R br and an inner h in and an outer h out scalelength,which is obtained by?tting two1D exponentials to the pro?les.

3.Results

A surprising result is that sample galaxies,although morphologically selected to look as similar as our simple model disk,frequently show signi?cant deviations from the input model.Some objects exhibit a huge low surface brightness enve-lope(e.g.ESO572-044,)or a rather perturbed outer component(e.g.UGCA250). In another case(e.g.ESO443-042)a strong bar hampers a reasonable model?t-ting.We?nd that three galaxies(NGC3390,NGC3717,NGC4696C),classi?ed as Sb,do not show a truncation feature in their radial surface brightness pro?les, but rather an S0-like outer component.For the remaining galaxies we derive a mean value of R co/h co=3.5±0.8and con?rm the suspected coupling of the two parameters,cut-o?radius and associated scalelength,for the sharply-truncated model.This is,in combination to a slightly di?erent way in identifying the cut-o?radius,the reason for the signi?cantly deviant results of van der Kruit& Searle(1982)and Pohlen et al.(2000a).

Radial structure of galactic stellar disks3

Figure2.Left panel:Distribution of the distance independent ratio

of R br to h in with Hubble type T(37galaxies).Right panel:The dis-

tribution of cut-o?radii versus maximum rotation velocity v rot above

square and below cross D=50Mpc(52galaxies).

The main result,however,is that most galaxies(>60%)are best?tted with

a two-slope or smoothly-truncated model.Fig.1clearly shows that the best ?tting sharply-truncated model does not?t well for the inner pro?les,which

exhibit more likely a two-slope behaviour.The second slope is well described by

another exponential decline.We derive for the distance independent ratio of the break radius to the inner scalelength a value of R br/h in=2.5±0.8(cf.Fig.2)

with mean values of7.6kpc for the inner and1.9kpc for the outer scalelength.

The mean extrapolated surface brightness of this break on the major axis is μbr=23.4±0.6V-mag/2′′andμbr=22.6±0.6R-mag/2′′.Additionally,35% of the galaxies are also well?tted at higher z-pro?les with the two-slope struc-ture.We want to emphasise that the pro?les show a rather sharp break,but are

not sharply truncated,implying that beyond the break we still?nd a remain-

ing disk.These two-slope pro?les could not be produced by any conventional dust distribution as simulations have shown.The various characteristic param-eters(e.g.R co/h co,R co/h∞,R br/h in)do not correlate with the Hubble type, whereas plotting R co and R br in linear units versus the maximum rotation veloc-ity reveals an unusual distribution(cf.Fig.2).A general trend is expected since the faster the galaxy rotates the more mass it has and therefore its size will also be larger.However,there is a surprisingly sharp limit in the size-velocity rela-tion apparent.Above a well de?ned diagonal line no galaxies are found,whereas there is no similar lower boundary for the expected diagonal.This would imply the existence of a maximal possible size for a given rotational velocity.

The crucial experiment to con?rm this two-slope structure is to observe the same behaviour for face-on galaxies.We have used the2.2m telescope at Calar Alto with CAFOS and obtained images(t exp≈180min)of three face-on galaxies, chosen to be as circular as possible and therefore intrinsically face-on and not of early type.Observations are made using an e?cient rectangular R-band?lter (R¨o ser R,RR)achieving a reliable photometry down toμRR=28.0mag/2′′, equivalent toμJR=28.4mag/2′′in Johnson R.We?nd that the azimuthally averaged pro?les as well as pro?les from individual sectors exhibit a similar two-slope structure(cf.Fig.3).The break radius occurs at the same surface brightness level(μbr=24.9±0.6JR-mag/2′′)as compared to the edge-on case, if the line of sight integration is taken into account.However,we derive a mean

4M.Pohlen,R.-J.Dettmar,R.L¨u tticke,&G.Aronica

Figure3.Azimuthally averaged radial surface brightness pro?les

(solid lines)of NGC5923(left panel)and UGC9837(right panel).

Plotted are the four90?-segments(thin lines)and the360?(thick line)

combined pro?le.

value for the ratio of the break radius to the inner scalelength R br/h in=3.9±0.7 (individually:4.3,4.2,and3.1)for the three galaxies.This does not?t well to the edge-on result and we probably still have a problem to?nd a good scalelength comparison for the edge-on and face-on case.However,the neglected dust and the applied1D?tting of the individual scalelengths in the edge-on case both tend to increase the measured scalelength and therefore decrease R br/h in.

4.Outlook

After this extensive optical imaging campaign the physical nature of disk trun-cations is still unknown and will be approached in the next step.We are still lacking of a detailed analysis to prove that these truncations in the optical light pro?les are also present in the mass distribution.In addition,one has to check for possible environmental e?ects to address a tidal truncation scenario.We will observe the radial molecular gas distribution—as the reservoir for star-formation—of galaxies with known optical truncations to?nd a correlation between the truncations and star-formation.A similar approach will be per-formed by a comparison study of the truncated optical pro?les with Hα-pro?les tracing the actual star-formation.

References

Ferguson,A.M.N.,&Clark,C.J.2001,MNRAS,325,781

Kennicutt,R.C.,Jr.1989,ApJ,344,685

Pohlen,M.,Dettmar,R.-J.,&L¨u tticke,R.2000a,A&A,357,L1

Pohlen,M.,Dettmar,R.-J.,L¨u tticke,R.et al.2000b,A&AS,144,405 Pohlen,M.,2001,PhD Thesis,Ruhr-University Bochum,Germany

van der Kruit,P.C.1979,A&AS,38,15

van der Kruit,P.C.1987,A&A,173,59

van der Kruit,P.C.,&Searle,L.1982,A&A,110,61