Constraints on secular evolution in unbarred spiral galaxies: understanding bulge and disk formation

April 23rd 2012 July 10th 2012 Marja Kristin Seidel Jesús Falcón - Barroso Instituto de Astrofísica de Canarias www.iac.es/project/traces Constraints on secular evolution in unbarred spiral galaxies: understanding bulge and disk formation disentangling disk heating agents April 23rd 2012 July 10th 2012 Marja Kristin Seidel Jesús Falcón - Barroso Instituto de Astrofísica de Canarias www.iac.es/project/traces [email protected]

Introduction: Disk heating

More than half of the stellar mass in the local Universe is found in disk galaxies (e.g. Driver et al. 2007) but we are far from understanding them.

Especially: What is driving the heating of the disk?

Investigate 3D-distribution of stellar velocity dispersions

ellipsoid with axes ratios: σz/σR (and σϕ/σR) σz σR Possible disk heating agents: • encounters with giant molecular clouds (GMCs) • scattering by dark halo objects or globular clusters • perturbation by spiral structure • perturbation by stellar bars talk on: www.iac.es/project/traces • dissolution of young stellar clusters • disturbances by satellite galaxies or minor mergers (e.g. from Spitzer & Schwarzschild, 1951 to Saha et al. 2010) [email protected]

Introduction: Disk heating

Model predictions • encounters with GMCs (Sellwood, 2008): 3D agent

• perturbation by spiral structure (Jenkins & Binney, 1990): σR

Age • radial migration (e.g. Roskar et al. 2008):

σR increases with age

Metallicity • for an existing metallicity gradient (e.g. Sánchez-Blázquez, 2009)* :

σz/σR increase with metallicity

talk on: www.iac.es/project/traces But: in the solar neighborhood little evidence for the age-metallicity-relation (e.g. Feltzing et al., 2001) *still present even with satellites! [email protected]

Introduction: Our study

Our study focuses on:

• 6 disk galaxies across the Hubble sequence

• obtaining the ages and metallicities in different regions of the galaxies via full-spectrum fitting techniques

• relating these stellar population parameters with earlier

kinematic results, i.e. σz/σR and the individual values (Shapiro & Gerssen 2003 and 2012)

talk on: www.iac.es/project/traces [email protected]

Sample Disk-heating (Shapiro & Gerssen 2003 and 2012)

credit: HST/WFPC2 image

spectra with resolutions of ~ 30 km s-1 and ~ 23 km s-1

NGC 2280 (Scd) NTT NGC 3810 (Sc) NGC 4030 (Sbc) NGC 1068 (Sb) KPNO NGC 2775 (Sa/Sab) NGC 2460 (Sa) different regions from radial surface brightness profiles and disk scale lengths: center bulge [trans] disk

talk on: www.iac.es/project/traces Note: all our galaxies show central sigma drops! [email protected] Methods pPXF (Cappellari & Emsellem, 2004) ; Gandalf (Sarzi et al. 2006) STARLIGHT (Cid Fernandes, 2007)

gal_3810mj_cor.fits [Bin 1 ; x = -15.0 ; SN= 15.5] 12 Starlight + Miles models NGC 3810 (Sc) DISKV= 1062.6, != 2.0 km/s (Sánchez-Blazquez et al.2006) gal_3810mj_cor.fits [Bin 1 ; x = -15.0 ; SN= 15.5] 10 12 V= 1062.6, != 2.0 km/s 1.2 8 10 pPXF 6 8 40 1 4 6

2 4 0 500 1000 1500 20 0.8 2 0 500 1000 1500 30 Gandalf 0 0.6 30 +ELODIE 20 (Prugniel et Intensity [arb.units] counts al. 2007) 20 10 0.4 40 counts 10 0

0 500 1000 1500 0 λpixels [Å] 20 0.2 0 500 1000 1500 pixels 0

For all three 4861.33 4958.91 regions 5006.84 in major 5175.36 and 5270.00 minor 5335.00 5406.00 axes: " H [OIII] [OIII] Mgb Fe5270 Fe5335 Fe5406 0 -0.2 10 talk on: www.iac.es/project/traces

4861.33 4958.91 5006.84 5175.36 5270.00 5335.00 ages 5406.00 "

mass &H luminosity[OIII] [OIII] weightedMgb Fe5270 Fe5335 Fe5406 4800 5000 5200 5400 0 5 10 15 8 metallicities 10 6 8

Counts Analyze4 the stellar populations! 6 2

Counts 4 0 4800 5000 5200 5400 2 Restwavelength [Angstroms]

0 4800 5000 5200 5400 Restwavelength [Angstroms] How can stellar populations help us to understand

• secular evolution in spirals?

• disk heating processes? [email protected]

Preliminary results: Ages and [Fe/H]

Luminosity weighted Mass weighted

age [Gyr]

[Fe / H]

repr. error 250 center bulge disk center bulge disk Sa Sab Young populations dominate the luminosity weighted age (e.g. Serra & Trager, 2006) talk on: www.iac.es/project/traces Sb Sbc We mostly confirm inside-out growth scenario (e.g. Muños-Mateos, 2007) 200 Sc In our sample: late types show stronger [Fe/H] gradients than the early types Scd (adding to MacArthur et al., 2009) R

! 150

100

50 0 5 10 15 20 age [Gyr] [email protected]

Preliminary results: Ages and [Fe/H]

Luminosity weighted Mass weighted

age [Gyr]

[Fe / H]

repr. error 250 center bulge disk center bulge disk Sa Sab Young populations dominate the luminosity weighted age (e.g. Serra & Trager, 2006) talk on: www.iac.es/project/traces Sb Sbc We mostly confirm inside-out growth scenario (e.g. Muños-Mateos, 2007) 200 Sc In our sample: late types show stronger [Fe/H] gradients than the early types Scd (adding to MacArthur et al., 2009) R

! 150

100

50 0 5 10 15 20 age [Gyr] [email protected]

Preliminary results: Disk heating

With our stellar population results: 6 J.Gerssen & K. Shapiro Griffin is it possible to relate these two findings?Evolutionary stellar population synthesis 1659

Z B-V

talk on: www.iac.es/project/traces Figure 4. Velocity ellipsoid ratio σz/σR as a function of galactic Figure 5. Velocity ellipsoid ratio σz/σR as a function of type (Hubble stage T). The solid points show the results that inclination- and extinction-corrected galaxy colour (available in age [Gyr] Figure 19. We plot the broad-band B V colour derived from the SSP we obtained previously (Shapiro et al. 2003). The results that we HyperLeda for 7 of our 8 spirals and the fast-rotating E/S0s). The − derive in this paper for two late type spirals are shown as the black points indicate our spiral galaxy data, and the red squaresSEDs for different ages and metallicities (as indicated within the panel) and Gerssen & Shapiro, 2012 the KroupaVazdekis universal IMF with et the al., zero-point 2010 set with two different Vega Figure 17. Variation of Balmer line-strength indices with metallicity pre- open circles. Horizontal errors represent the uncertainty inherent are the data of Cappellari et al. (2007) for fast-rotating E andspectra in black (very thin) and red (thicker) (see the text for details). The dicted by our single stellar population models of 10 Gyr and Kroupa uni- thickest green line represents the photometric colour computed by Vazdekis in galaxy classification (Naim et al. 1995). The filled square is the S0 galaxies.versal The IMF. linear Different fit line (dashed styles show line) predictions is to for theindices spiral measured gala at xies et al. (1996), as updated in this work, which is based on extensive empirical value in the Solar Neighbourhood derived from Hipparcos data only. The probabilitydifferent spectral of resolutions, no correlation, as in Fig. 14. null hypothesis H , is one 0 photometric stellar libraries. (Dehnen & Binney 1998). percent.

of 4300 Å are significantly larger than for the redder part of the units (IFUs) to observe velocity dispersions in disk galaxies. “fast rotators,” are bulge-dominated galaxies that neverthe-spectra∼ for all the clusters. We do not discard possible effects due to This is observationally more efficient than obtaining long-slit less contain a significant disk component (Kuntschner et al.the fact that the Milky Way GCs show oxygen-enhanced abundance spectra, one at a time, along two (or more) position angles. ratios for low metallicities. In fact, Cassisi et al. (2004) have shown 2006; Krajnovi´cet al. 2008). It is therefore interesting tothat such effects become relevant for the spectral ranges covered by Additionally, since IFUs uniformly sample velocity disper- investigate how the disks of these early-type galaxies arethe B, and mostly U, broad-band filters. Therefore, an appropriate sions along both azimuth and radius across the disk, the related to those in spiral galaxies. modelling for these clusters requires working with such α-enhanced assumption of the epicycle approximation, employed here stellar evolutionary isochrones as well as the use of stellar spectra Cappellari et al. (2007) have used axisymmetricwith a similar abundance pattern. In addition, we also require to for long-slit data, can be relaxed. Noordermeer et al. (2008) Schwarzschild dynamical models to extract the three-use specific stellar spectra with CN-strong absorption features and use the PPAK IFU to explore velocity dispersions in disk dimensional orbital structure of a subsample of theCN-enhanced isochrones. galaxies and to constrain the shape of the velocity ellipsoid In Fig. 21, we show our SED fit to the integrated spectrum of SAURON galaxies and measure the shape of their velocitythe standard open cluster M67 (Schiavon, Caldwell & Rose 2004). in one system, NGC 2985. Their measurement of σz/σR ellipsoids. These anisotropy measurements are luminosity-Unlike for the GCs of Fig. 20, these authors obtained the integrated ≈ 0.7 is consistent with our previous result for this sys- weighted, giving more weight to the high-density equatorialspectrum for M67 by co-adding individual spectra of cluster mem- tem (0.75 ± 0.09; Gerssen et al. 2000) and confirms that the bers, weighted according to their luminosities and relative numbers. plane, and volume-averaged, giving more weight to largerThe age and metallicity obtained are in good agreement with these epicycle theory we assume is indeed applicable in this galaxy. radii; as aFigure result, 18. Comparison the global of the Hβ anisotropiesindex computed on (see the basis table of the 2 ofauthors (see also Schiavon 2007) and with our isochrone fitting re- In a series of conference proceedings Westfall et al. empirical fitting functions of Worthey et al. (1994), using the models of sults (Vazdekis et al. 1996) within the uncertainties. The residuals CappellariVazdekis et al. et 2007) al. (1996), are as updated dominated in this work and by transformed the disks to the andLIS- are only show a low-frequency pattern, which is related to differences (2008, 2010) present their ongoing work with the SparsePAK therefore8.4 comparable Å system following to Table our A2, with measurements the index measured on the in SSP later-type SEDs, and PPAK IFUs to constrain the velocity ellipsoid shape once smoothed to the LIS-8.4 Å resolution. For the lowest metallicity, we in the flux calibration. However, the higher frequency residuals are galaxies (Cappellarionly plot the results et for SSPs al. older private than 10 Gyr communication). according to our qualitative Sincenegligible. It is particularly interesting that, unlike in Fig. 20, we and its radial variation in a sample of spiral galaxies. These Hubble T-typeanalysis shown is less in the second meaningful panel of Fig. in 6. However, early-type for the sake galaxies of do not see the residuals bluewards of 4300 Å. This shows that all authors find a strong dependence on their modelled veloc- the discussion we also show the predictions based on the Lick/IDS fitting relevant spectral types are present in the MILES library. Therefore, than in spirals,functions for we [M/H] cannot1.7, even add though these we do not points consider them directly to be of tothe residuals obtained for Galactic GCs can be attributed to the = − ity ellipsoid shape with their measurement techniques and Figure 4. Insteadgood quality we because use of the galaxy low numberB − of starsV colour at those metallicities. as a proxy fornon-solar abundance ratios present in these clusters. However, we assumptions; they present and describe an analysis compa- morphological type and plot this quantity against σz/σR in C 2010 The Authors. Journal compilation C 2010 RAS, MNRAS 404, 1639–1671 rable to ours for a single galaxy, NGC 3982 (Sb), for which Figure 5 for# both the SAURON sample# and our sample of they find σz/σR =0.31 − 0.73 over a radial range of 1 − 2 spiral galaxies. Colours are taken from HyperLeda2 and are photometric scalelengths, broadly consistent with our results listed for our sample in Table 4. for galaxies of similar Hubble type. These authors also find The combined data span the Hubble sequence from E some evidence for variation in the velocity ellipsoid ratio to Scd and show the strong correlation between velocity el- with radius, indicating the potential of IFU data in future lipsoid ratio and galaxy colour, as expected from Figure 4, studies of the velocity ellipsoid in external galaxies. given the known relationship between galaxy colour and Hubble type. In Figure 5, we find a continuous trend of in- 4.4 The Velocity Ellipsoid across the Hubble creasing anisotropy in bluer galaxies. Moreover, we find that Sequence the anisotropies of the E/S0 galaxies overlap with those of the earliest-type spirals in our sample. However, in the bluest The SAURON team has used the IFU of the same name to study the kinematics of elliptical and lenticular galaxies and have shown that many early-type galaxies, the so-called 2 http://leda.univ-lyon1.fr/

!c 0000 RAS, MNRAS 000, 000–000 [email protected]

Preliminary results: σz/σR

So far, only tentative correlations for both, [Fe/H] being stronger.

1.4 1.4 Sa Sa H0 = 80% H0 = 20% Sab Sab 1.21.2 1.2 Sb Sb Sbc Sbc

] Sc Sc -1 1.0 1.0 Scd Scd

[ km s 0.80.8 0.8

R R ! ! R / / z z σ ! ! /

z 0.60.6 0.6 σ

0.40.4 0.4

0.20.2 0.2 250 0.00.0 0.0 Sa 0 5 10 15 -1.2 -1.020 -0.8 -0.6 -0.4 -0.2 -0.0 0.2 Sab age [Gyr]age [Gyr] [MgFe][Fe/H] Sb talk on: www.iac.es/project/traces Sbc 200 Sc Scd R

! 150

100

50 0 5 10 15 20 age [Gyr] [email protected]

Preliminary results: σR

Best correlation obtained for σR with age: 250 H0 = 30% Sa Sab Sb Sbc 200 Sc Scd ] -1 R

! 150 [ km s

R σ

100

250 50 Sa 0 5 10 15 20 Sab ageage [Gyr] [Gyr] Sb talk on: www.iac.es/project/traces Sbc Consistent with simulated predictions (Roskar et al. 2008a, 2008b): 200 Sc if radial migration of stars is present, one expects an increase of σR with age Scd R

! 150

100

50 0 5 10 15 20 age [Gyr] L66 ROSˇKAR ET AL. Vol. 675

Each component is modeled by10 6 particles; the dark matter halo is composed of two shells, with the inner halo of 9 # 5 6 10 particles (each of mass 1 # 10 M,) extending to 200 kpc and an outer halo of1 #∼10 5 particles (each of mass 3.5 6 ∼ # 10 M,) beyond. All gas particles initially have a mass of [email protected] 5 1.4 # 10 M,. We use a softening length of 100 pc for all dark matter particles and 50 pc for baryonic particles. We adopt Discussion: σR the best-fit values from Stinson et al. (2006) for the parameters governing the physics of star formation and feedback. Our cooling prescriptions do not account for effects of UV back- Best correlation obtained for σR with age: ground or metal line cooling. The global criteria for SF are 250 Ϫ3 H0 = 30% Sa 1 ! Sab that a gas particle has to haven 0.1 cm , T 15,000 K and Sb Sbc be a part of a converging flow; efficiency of star formation is 200 L66Sc ROSˇKAR ET AL. Vol. 675 Scd 0.05, i.e., 5% of gas eligible to form stars is converted into ] -1 stars per dynamical time. Star6 particles form with an initial R Each component is modeled by10 particles; the dark matter ! 150 mass of 1/3 gas particle mass, which at our resolution corre- [ km s halo is composed of two4 shells, with the inner halo of 9 #

R sponds5 to 4.6 # 10 M,. A gas6 particle may spawn multiple

σ 10 particles (each of mass 1 # 10 M,) extending to 200 100 kpcstar and particles an outer but halo to of avoid1 #∼10 gas5 particles particles (each of unreasonably of mass 3.5 small mass,6 the minimum gas particle mass is restricted to∼ 1/5 of the # 10 M,) beyond. All gas particles initially have a mass of 50 1.4initial# 10 5 mass.M . We The use simulation a softening is evolved length of with 100 the pc forparallel all SPH 0 5 10 15 20 , age [Gyr] darkcode matter GASOLINE particles and (Wadsley 50 pc for et baryonic al. 2004) particles. for 10 We Gyr. adopt age [Gyr] 57 250 the best-fitWe have values also from performed Stinson et simulations al. (2006) for with the parameters10 and 10 par- Sa governingticles per the component, physics of thus star formation bracketing and our feedback. fiducial run. Our While cooling prescriptions do not account for effects of UV back- Sab Fig. 1.—Azimuthally averaged properties of the disk attalk 1, 5, on: and www.iac.es/project/traces 10 Gyr. the details of the gas cooling are somewhat resolution-depen- Sb ground or metal line cooling. The global criteria for SF are Consistent with simulated predictionsThe top (Roskar panels et show al. 2008a, the stellar 2008b) surface: density profiles, with the dotted lines dent (Kaufmann et al. 2006), resulting in morphological dif- Sbc representing double exponential fits. The two exponential components were that a gas particle has to haven 1 0.1 cmϪ3, T ! 15,000 K and Sc if radial migration of stars is present, one expects an increase of σR with age ferences, we find convergence in the modeling of overall halo 200 fit simultaneously to the profile, excluding the innermost and outermost regions. Scd becooling, a part of star a converging formation flow; and efficiencyits dependence of star on formation gas surface is den- The point of intersection of the two exponentials is taken as the break radius. 0.05, i.e., 5% of gas eligible to form stars is converted into In all panels, the vertical lines indicate the location of the break. The second sity, and disk dynamics. We therefore consider the conclusions starsand per predictions dynamical of time. this Star Letter particles numerically form with robust. an The initial10 6 par- row shows the surface density of cool gas. In the third row, we show the SFR mass of 1/3 gas particle mass, which at our resolution corre- density calculated from the stellar mass formed in the previous 10 Myr. The ticle resolution4 used here represents a compromise between

R mean stellar age (normalized by time of output) as a function of radius is sponds to 4.6 # 10 M,. A gas particle may spawn multiple

! 150 adequate statistics for detailed analysis in the outer disk and shown in the bottom row. star particles but to avoid gas particles of unreasonably small mass,computational the minimum cost gas for particle 10 Gyr mass of is evolution. restricted to 1/5 of the Our simulation should be thought of as modeling disk for- formation which is caused in our simulation by a rapid decrease initial mass. The simulation is evolved with the parallel SPH mation and evolution from the cooling of hot gas after the last in the surface density of cool gas; (3) the outer disk is populated code GASOLINE (Wadsley et al. 2004) for 10 Gyr. gas-richWe have majoralso performed merger, simulations as suggested with by10 cosmological57 and 10 par- simula- 100 by stars that have migrated, on nearly circular orbits, from the ticlestions per (Brook component, et al. thus 2004, bracketing 2007). our Although fiducial werun. make While use of inner disk, and consequently the break is associated with a Fig. 1.—Azimuthally averaged properties of the disk at 1, 5, and 10 Gyr. thesimplifications details of the gas such cooling as an initiallyare somewhat spherical resolution-depen- distribution of mat- sharp change in the radial mean stellar age profile; and (4) The top panels show the stellar surface density profiles, with the dotted lines dentter, (Kaufmann a lack of et halo al. 2006), substructure, resulting the in morphological ignoring of subsequent dif- breakrepresenting parameters double agree exponential with fits. current The two observations. exponential components were ferences,accretion we of find dark convergence matter and in baryons, the modeling and of an overall initial gashalo density 50 fit simultaneously to the profile, excluding the innermost and outermost regions. cooling, star formation and its dependence on gas surface den- The point of intersection of the two exponentials is taken as the break radius. profile that mimics that of the dark matter, studying the ide- 0 5 10 15 20 In all panels, the2. verticalSIMULATION lines indicate METHODOLOGY the location of the break. The second sity,alized and disk isolated dynamics. case allows We therefore us to analyze consider in the detail conclusions the important and predictions of this Letter numerically robust. The10 6 par- age [Gyr] Resolvedrow shows the stellar surface population density of cool data gas. In in the disk third outskirts, row, we show which the SFR are dynamical processes affecting the evolution of a massive iso- density calculated from the stellar mass formed in the previous 10 Myr. The ticle resolution used here represents a compromise between nowmean becoming stellar age available (normalized (e.g., by time Ferguson of output) as et a al. function 2007; of Barker radius is et adequatelated disk. statistics The for lessons detailed we analysis learn from in the this outer idealized disk and case will al.shown 2007; in de the bottomJong etrow. al. 2007), provide strong constraints on computationallater be applied cost for to 10 galaxies Gyr of evolution. evolved in a full cosmological theories of break formation. Therefore, the inclusion of SF and contextOur simulation (R. Ros shouldˇkar et be al., thought in preparation). of as modeling disk for- formation which is caused in our simulation by a rapid decrease feedback is required to assess break formation models. For this mation and evolution from the cooling of hot gas after the last in the surface density of cool gas; (3) the outer disk is populated reason we have run simulations of gas cooling, collapsing, and gas-rich major merger, as suggested by cosmological simula- by stars that have migrated, on nearly circular orbits, from the 3. BREAK FORMATION AND THE OUTER DISK forming stars inside a live dark matter halo within which the tions (Brook et al. 2004, 2007). Although we make use of inner disk, and consequently the break is associated with a simplifications such as an initially spherical distribution of mat- gassharp is initially change in in hydrostatic the radial mean equilibrium. stellar age This profile; has the and further (4) In Figure 1, the break is already evident as soon as a stable ter, a lack of halo substructure, the ignoring of subsequent advantagebreak parameters of making agree no with assumptions current observations. about the angular mo- disk forms at 1 Gyr, moving outward as the disk grows and mentum distribution within the disk, which can strongly affect accretionpersisting of dark throughout matter and the baryons, simulation. and an A initial sharp gas drop density in the local profile that mimics that of the dark matter, studying the ide- its subsequent evolution2. SIMULATION (Debattista METHODOLOGY et al. 2006). SFR is always present at the break radius. The drop in SFR is We construct initial conditions as in Kaufmann et al. (2007). alizednot isolated due to a case volume allows density us to analyze threshold in detail for the star important formation, but Resolved stellar population data in disk outskirts, which are dynamical processes affecting the evolution of a massive iso- The initial system consists of a virialized spherical NFW dark is instead associated with a rapid decrease in the gas surface now becoming available (e.g., Ferguson et al. 2007; Barker et lated disk. The lessons we learn from this idealized case will matteral. 2007; halo de (Navarro Jong et et al. al. 2007), 1997) provide and an strong embedded constraints spherical on laterdensity be applied (the star to formation galaxies evolved follows in a aKennicutt-Schmidt full cosmological law at hottheories baryonic of break component formation. containing Therefore, 10% the of inclusion the total of mass SF and and contextall times) (R. Ros andˇkar a etcorresponding al., in preparation). sharp increase in the Toomre Q followingfeedback the is required same density to assess distribution, break formation which models. at the For end this of parameter. We verified that the break is not seeded by our star thereason simulation we have yields run simulations a disk mass of fraction gas cooling, of collapsing,5%. The and mass formation recipe by running several simulations with different 12 ∼ 3. BREAK FORMATION AND THE OUTER DISK withinforming the virialstars inside radius a is live10 darkM, matter. A temperature halo within which gradient the in values of the threshold density and found that the location of thegas gas is component initially in hydrostatic ensures an equilibrium. initial gas pressure This has theequilibrium further theIn Figure break 1, did the not break depend is already on the evident particular as soon value as a stable used. Since foradvantage an adiabatic of making equation no of assumptions state. Velocities about the of angular gas particles mo- diskdensity forms is at inversely 1 Gyr, moving proportional outward as to the radius, disk the grows cooling and time arementum initialized distribution according within to a cosmologicallythe disk, which can motivated strongly specific affect persistingincreases throughout outward. the By simulation. construction, A sharp the drop angular in the momentum local is angularits subsequent momentum evolution distribution (Debattista with etj al.r 2006). and an overall spin SFRdirectly is always proportional present at the to cylindricalbreak radius. radius, The drop which in SFR means is that parameterWe constructl p ( j/ initialG)(FE conditionsF/M 3) 1/2 p as0.039 in Kaufmann∝ (Bullock et et al. al. (2007). 2001). nothigher due to angular a volume momentum density threshold material for will star formation,take longer but to cool. The initial system consists of a virialized spherical NFW dark is instead associated with a rapid decrease in the gas surface matter halo (Navarro et al. 1997) and an embedded spherical density (the star formation follows a Kennicutt-Schmidt law at hot baryonic component containing 10% of the total mass and all times) and a corresponding sharp increase in the Toomre Q following the same density distribution, which at the end of parameter. We verified that the break is not seeded by our star the simulation yields a disk mass fraction of 5%. The mass formation recipe by running several simulations with different 12 ∼ within the virial radius is 10 M,. A temperature gradient in values of the threshold density and found that the location of the gas component ensures an initial gas pressure equilibrium the break did not depend on the particular value used. Since for an adiabatic equation of state. Velocities of gas particles density is inversely proportional to radius, the cooling time are initialized according to a cosmologically motivated specific increases outward. By construction, the angular momentum is angular momentum distribution withj r and an overall spin directly proportional to cylindrical radius, which means that parameterl p ( j/G)(FEF/M 3) 1/2 p 0.039∝ (Bullock et al. 2001). higher angular momentum material will take longer to cool. 10 J.Gerssen & K. Shapiro Griffin

[email protected]

Discussion: Spiral structure - σR Figure 7. The velocity ellipsoid shape and magnitudes as a function of the H2 gas surface density. Gas densities are estimated from the CO measurements of Young et al. (1995) and are computed as average values over the radii used in our kinematic analysis. Left: The velocity ellipsoid shapes are not correlated with the molecular gas surface density. Middle: There is a hint that the vertical component of the velocity dispersion is correlated with Σ , but the scatter is too large to state this conclusively. Right: The radial component H2 Transcient spiral arms (Sellwood&Binney, 2002), no increase in σz, but increase in σR increases with molecular gas density. Comparison with arm-class (Gerssen & Shapiro, 2012): best for σR

250 H0 = 30% Sa Sab Sb Sbc 200 Sc Scd ] -1 R

! 150 [ km s

R σ 100

250 50 0 5 10 15 20 Sa age [Gyr][Gyr] Figure 8. The velocity ellipsoid shape and magnitudes as a function of armSab class, as defined in Elmegreen & Elmegreen (1987) to talk on: www.iac.es/project/traces quantify the orderliness of spiral structure from flocculent (class 1) to grand-designSb (class 12). Note that there is no arm-class 10 and 11 BUT: trend for σz with both age, metallicity and arm-class found (cf., Figure 1 in Elmegreen & Elmegreen 1987). Left: The velocity ellipsoid shapeSbc is not correlated with arm class. Middle: The vertical 200 magnitude of the ellipsoid decreases with arm class. Right: There is a clear trendSc between arm class and radial however, component not, as as expected strong as for σR . from the Toomre Q criterion, see text. Note that for plotting purposes we have addedScd small offsets to galaxies with the same arm class. here. This research has made use of the Hyperleda database Kuntschner H., McDermid R. M., Peletier R. F., Sarzi M., (http://leda.univ-lyon1.fr) and of the NASA/IPAC Extra- van den Bosch R. C. E., van de Ven G., 2007, MNRAS, galactic Database, which is operated by the Jet Propulsion 379, 418 R Laboratory, California Institute of Technology, under con- ! 150 Carlberg R. G., Dawson P. C., Hsu T., Vandenberg D. A., tract with the National Aeronautics and Space Administra- 1985, ApJ, 294, 674 tion. Casetti-Dinescu D. I., Girard T. M., Korchagin V. I., van Altena W. F., 2011, ApJ, 728, 7 Comer´onS., Knapen J. H., Beckman J. E., 2008, A&A, REFERENCES 100 485, 695 Barbanis B., Woltjer L., 1967, ApJ, 150, 461 Dehnen W., Binney J. J., 1998, MNRAS, 298, 387 Benson A. J., Lacey C. G., Frenk C. S., Baugh C. M., Cole Driver S. P., Allen P. D., Liske J., Graham A. W., 2007, S., 2004, MNRAS, 351, 1215 ApJ, 657, L85 Binney J., 2012, ArXiv e-prints Elmegreen D. M., Elmegreen B. G., 1987, ApJ, 314, 3 Binney J., Dehnen W., Bertelli G., 2000, MNRAS, 318, 658 Eskridge P. B., Frogel J. A., Pogge R. W., Quillen A. C., 50 Cappellari M., Emsellem E., 2004, PASP, 116, 138 Berlind A. A., Davies R. L., DePoy D. L., Gilbert K. M., Cappellari M., Emsellem E., Bacon R., Bureau M., Davies Houdashelt M. L., Kuchinski L. E., Ram´ırezS. V., Sell- 0 R. L., de5 Zeeuw P. T., Falc´on-Barroso10 J., Krajnovi´cD.,15 gren K., Stutz20 A., Terndrup D. M., Tiede G. P., 2002, age [Gyr] !c 0000 RAS, MNRAS 000, 000–000 [email protected]

Discussion: GMCs - σz

GMCs (Sellwood, 2008), increase in σz and σR

10 J.Gerssen & K. Shapiro GriffiComparisonn with H2 gas surface density (Young et al., 1995):σz

250 Sa H0 = 44% Sab Sb 200 Sbc Sc Scd ]

-1 150 z ! [ km s 100 z σ

50

0 250 0 5 10 15 20 age [Gyr] Sa age [Gyr] Sab Figure 7. The velocity ellipsoid shape and magnitudes as a function of the H2 gas surface density. Gas densities are estimated from the talk on: www.iac.es/project/traces CO measurements of Young et al.Sb (1995) and are computed as average values over the radii used in our kinematic analysis. Left: The velocity ellipsoid shapes are notSbc correlated with the molecSurprisinglyular gas surface not a density. very goodMiddle: correlationThere is a with hint σ thatz the vertical component 200 of the velocity dispersion is correlatedSc with ΣH2 , but the scatter is too large to state this conclusively. Right: The radial component increases with molecular gas density.Scd R

! 150

100

50 Figure 8. The velocity ellipsoid shape and magnitudes as a function of arm class, as defined in Elmegreen & Elmegreen (1987) to 0 5 10 quantify the orderliness15 of spiral structure20 from flocculent (class 1) to grand-design (class 12). Note that there is no arm-class 10 and 11 (cf., Figure 1 in Elmegreen & Elmegreen 1987). Left: The velocity ellipsoid shape is not correlated with arm class. Middle: The vertical age [Gyr] magnitude of the ellipsoid decreases with arm class. Right: There is a clear trend between arm class and radial component, as expected from the Toomre Q criterion, see text. Note that for plotting purposes we have added small offsets to galaxies with the same arm class. here. This research has made use of the Hyperleda database Kuntschner H., McDermid R. M., Peletier R. F., Sarzi M., (http://leda.univ-lyon1.fr) and of the NASA/IPAC Extra- van den Bosch R. C. E., van de Ven G., 2007, MNRAS, galactic Database, which is operated by the Jet Propulsion 379, 418 Laboratory, California Institute of Technology, under con- Carlberg R. G., Dawson P. C., Hsu T., Vandenberg D. A., tract with the National Aeronautics and Space Administra- 1985, ApJ, 294, 674 tion. Casetti-Dinescu D. I., Girard T. M., Korchagin V. I., van Altena W. F., 2011, ApJ, 728, 7 Comer´onS., Knapen J. H., Beckman J. E., 2008, A&A, REFERENCES 485, 695 Barbanis B., Woltjer L., 1967, ApJ, 150, 461 Dehnen W., Binney J. J., 1998, MNRAS, 298, 387 Benson A. J., Lacey C. G., Frenk C. S., Baugh C. M., Cole Driver S. P., Allen P. D., Liske J., Graham A. W., 2007, S., 2004, MNRAS, 351, 1215 ApJ, 657, L85 Binney J., 2012, ArXiv e-prints Elmegreen D. M., Elmegreen B. G., 1987, ApJ, 314, 3 Binney J., Dehnen W., Bertelli G., 2000, MNRAS, 318, 658 Eskridge P. B., Frogel J. A., Pogge R. W., Quillen A. C., Cappellari M., Emsellem E., 2004, PASP, 116, 138 Berlind A. A., Davies R. L., DePoy D. L., Gilbert K. M., Cappellari M., Emsellem E., Bacon R., Bureau M., Davies Houdashelt M. L., Kuchinski L. E., Ram´ırezS. V., Sell- R. L., de Zeeuw P. T., Falc´on-Barroso J., Krajnovi´cD., gren K., Stutz A., Terndrup D. M., Tiede G. P., 2002,

!c 0000 RAS, MNRAS 000, 000–000 [email protected]

Discussion: GMCs - σR

GMCs (Sellwood, 2008), increase in σz and σR

10 J.Gerssen & K. Shapiro Griffin Comparison with H2 gas surface density (Young et al., 1995): best for σR

250 H0 = 30% Sa Sab Sb Sbc 200 Sc Scd ] -1 R

! 150 [ km s

R σ 100

250 50 0 5 10 15 20 Sa age [Gyr][Gyr] Sab Figure 7. The velocity ellipsoid shape and magnitudes as a function of the H2 gas surface density. Gas densities are estimated from the talk on: www.iac.es/project/traces CO measurements of Young et al. (1995) and are computed as average values overSb the radii used in our kinematic analysis. Left: The velocity ellipsoid shapes are not correlated with the molecular gas surface density.SbcMiddle: There is aCorrelation hint that the with vertical σR component better! Adding to spiral structure? 200 of the velocity dispersion is correlated with ΣH2 , but the scatter is too large toSc state this conclusively. Right: The radial component increases with molecular gas density. Scd R

! 150

100

50 Figure 8. The velocity ellipsoid shape and magnitudes as a function of arm class, as defined in Elmegreen & Elmegreen (1987) to 0 quantify the5 orderliness of spiral structure10 from flocculent (class15 1) to grand-design (class20 12). Note that there is no arm-class 10 and 11 (cf., Figure 1 in Elmegreen & Elmegreen 1987). Left: The velocity ellipsoid shape is not correlated with arm class. Middle: The vertical magnitude of the ellipsoid decreasesage with [Gyr] arm class. Right: There is a clear trend between arm class and radial component, as expected from the Toomre Q criterion, see text. Note that for plotting purposes we have added small offsets to galaxies with the same arm class. here. This research has made use of the Hyperleda database Kuntschner H., McDermid R. M., Peletier R. F., Sarzi M., (http://leda.univ-lyon1.fr) and of the NASA/IPAC Extra- van den Bosch R. C. E., van de Ven G., 2007, MNRAS, galactic Database, which is operated by the Jet Propulsion 379, 418 Laboratory, California Institute of Technology, under con- Carlberg R. G., Dawson P. C., Hsu T., Vandenberg D. A., tract with the National Aeronautics and Space Administra- 1985, ApJ, 294, 674 tion. Casetti-Dinescu D. I., Girard T. M., Korchagin V. I., van Altena W. F., 2011, ApJ, 728, 7 Comer´onS., Knapen J. H., Beckman J. E., 2008, A&A, REFERENCES 485, 695 Barbanis B., Woltjer L., 1967, ApJ, 150, 461 Dehnen W., Binney J. J., 1998, MNRAS, 298, 387 Benson A. J., Lacey C. G., Frenk C. S., Baugh C. M., Cole Driver S. P., Allen P. D., Liske J., Graham A. W., 2007, S., 2004, MNRAS, 351, 1215 ApJ, 657, L85 Binney J., 2012, ArXiv e-prints Elmegreen D. M., Elmegreen B. G., 1987, ApJ, 314, 3 Binney J., Dehnen W., Bertelli G., 2000, MNRAS, 318, 658 Eskridge P. B., Frogel J. A., Pogge R. W., Quillen A. C., Cappellari M., Emsellem E., 2004, PASP, 116, 138 Berlind A. A., Davies R. L., DePoy D. L., Gilbert K. M., Cappellari M., Emsellem E., Bacon R., Bureau M., Davies Houdashelt M. L., Kuchinski L. E., Ram´ırezS. V., Sell- R. L., de Zeeuw P. T., Falc´on-Barroso J., Krajnovi´cD., gren K., Stutz A., Terndrup D. M., Tiede G. P., 2002,

!c 0000 RAS, MNRAS 000, 000–000 [email protected]

Summary and Outlook

check figure Potential relation between disk heating agents and stellar ages and Z. Strongest suspect: spiral structure in addition with GMCs: Brad Gibson papers!!

1) best correlation for [Fe/H] with σz/σR

2) good correlation for age with σR

However, it will be interesting to check:

• truly radial dependencies within the disk to even better compare with model predictions

• separate distinct populations in our SFHs to better understand the interplay of bulge, disk and individual components talk on: www.iac.es/project/traces • obtain a larger sample of galaxies in order to increase our statistics and the reliability of our results [email protected] Thank you for listening

…any questions? - apart from this one...

talk on: www.iac.es/project/traces

! talk on: www.iac.es/project/traces [email protected]

Backup

MORE SLIDES TO ILLUSTRATE IF WANTED

talk on: www.iac.es/project/traces [email protected]

Backup 1990MNRAS.245..305J

talk on: www.iac.es/project/traces (Jenkins & Binney, 1990) L66 ROSˇKAR ET AL. Vol. 675

Each component is modeled by10 6 particles; the dark matter halo is composed of two shells, with the inner halo of 9 # 5 6 10 particles (each of mass 1 # 10 M,) extending to 200 kpc and an outer halo of1 #∼10 5 particles (each of mass 3.5 6 ∼ # 10 M,) beyond. All gas particles initially have a mass of 5 1.4 # 10 M,. We use a softening length of 100 pc for all dark matter particles and 50 pc for baryonic particles. We adopt the best-fit values from Stinson et al. (2006) for the parameters governing the physics of star formation and feedback. Our [email protected] cooling prescriptions do not account for effects of UV back- Discussion: Disk heating ground or metal line cooling. The global criteria for SF are 1 Ϫ3 ! 602 P. Sanchez-Bl´ azquez´ et al. that a gas particle has to haven 0.1 cm , T 15,000 K and be a part of a converging flow; efficiency of star formation is L66 ROSˇKAR ET AL. Vol. 675 Comparison with Roskar et al. 2008 and602 Sánchez-BlázquezP. Sanchez-Bl´ azquez´ et etal. al. 2009:0.05, i.e., 5% of gas eligible to form stars is converted into

stars per dynamical time. Star6 particles form with an initial Eachmass component of 1/3 gas is modeled particle by mass,10 which particles; at the our dark resolution matter corre- halo is composed of two4 shells, with the inner halo of 9 # sponds5 to 4.6 # 10 M,. A gas6 particle may spawn multiple 10 particles (each of mass 1 # 10 M,) extending to 200 kpcstar and particles an outer but halo to of avoid1 #∼10 gas5 particles particles (each of unreasonably of mass 3.5 small mass,6 the minimum gas particle mass is restricted to∼ 1/5 of the

Age [Gyr] # 10 M,) beyond. All gas particles initially have a mass of initial5 mass. The simulation is evolved with the parallel SPH 1.4 # 10 M,. We use a softening length of 100 pc for all darkcode matter GASOLINE particles and (Wadsley 50 pc for et baryonic al. 2004) particles. for 10 We Gyr. adopt the best-fitWe have values also from performed Stinson et simulations al. (2006) for with the parameters1057 and 10 par- governingticles per the component, physics of thus star formation bracketing and our feedback. fiducial run. Our While Figure 21. Star formation over the last 1 Gyr divided by the total mass of Fig. 1.—Azimuthally averaged properties of the disk at 1, 5, and 10 Gyr. coolingthe details prescriptions of thestars formedgas do not before cooling account this epoch are (the for so-called somewhat effects ‘birth rate’) of as resolution-depen- a UV function back- of Galactocentric radius. The top panels show the stellar surface density profiles, with the dotted lines grounddent or(Kaufmann metal lineFigure et cooling. 21. al.Star formation 2006), The over resulting the global last 1 Gyr criteriadivided in by morphological the fortotal mass SF of are dif- stars formed before this epoch (the so-calledϪ3 ‘birth rate’) as a function of representingConsistent double exponential with both fits. simulations, The two exponential but: components were thatferences, a gas particle we find hasGalactocentric convergence to have radius. radius.n 1 As0.1 can in be cm theseen, b modeling,increasesT ! 15,000 almostof linearly K overall and halo fit simultaneously to the profile, excluding the innermost and outermost regions. with radius, as expected for an inside-out formation scenario, until which can be the disk heating agent ? be a part of a convergingthe break radius, flow; where efficiency it reaches a plateau of and star then decreases. formation The is [Z/H] cooling, star formation and its dependence on gas surface den- The point of intersection of the two exponentials is taken as the break radius. Galactocentric radius. As can be seen, b increases almost linearly 0.05, i.e., 5% of gaswithincrease radius, eligible in the as expected error to bars forform at an the inside-out break stars radius formation is reflects converted scenario, the asymme- until into In all panels, the vertical lines indicate the location of the break. The second sity, and disktalk dynamics. trieson: in thewww.iac.es/project/traces age distribution We therefore of the stars beyond consider this radius the (recall conclusions stars per dynamicalthe break time. radius, Star where particles it reaches a plateau form and then with decreases. an The initial 6 row shows the surface density of cool gas. In the third row, we show the SFR and predictionsincreaseFig. of 5). this in the Letter error bars at numerically the break radius reflects robust. the asymme- The10 par- mass of 1/3 gas particleWe argue mass, that the U-shape which age at profile our is the resolution direct consequence corre- density calculated from the stellar mass formed in the previous 10 Myr. The tries in the age distribution of the stars beyond this radius (recall ticle resolutionFig.of4 used the 5). existence here of a break represents in the star formation a compromise density. If the star between sponds to 4.6 # 10 M,. A gas particle may spawn multiple mean stellar age (normalized by time of output) as a function of radiusR is [kpc] adequate statisticsformationWe argue for outside that detailed the the U-shape break had age analysis profilenot decreased is the in direct suddenly, the consequence outer the age disk and shown in the bottom row. star particles but toofgradient avoidthe existence would gas decrease, of a particles break or in remain the star of constant, formation unreasonably until density. the edge If the of star thesmall Figure 19. Mass-weighted azimuthally averagedmass, stellarcomputational age the and metallicity minimum costformationoptical gas disc. for particle outside This 10 is the supported Gyr mass break had of by is not the evolution. restricted decreased result of Bakos suddenly, to et 1/5al. the (2008) age of the gradients. Solid lines: theoretical gradient for the hypothetical case where gradientwho only would found decrease, the telltale or remainU-shaped constant, colour until profiles the for edge galaxies of the stars do not migrate from their birth place. Dash–dottedOur line: profile simulation mea- possessing should a Type II be profile. thought The galaxies of with as an essentially modeling pure disk for- Figure 19. Mass-weighted azimuthally averagedinitial stellar age mass. and metallicity The simulationoptical disc. This is is supported evolved by the with result of the Bakos parallel et al. (2008) SPH formation which is caused in our simulationsured by in a the rapid final time-step decrease of our simulation. Dotted line: profile measured gradients. Solid lines: theoretical gradient for the hypothetical case where whoexponential only found profile the within telltale their U-shaped sample colour showed profiles a plateau for galaxies(and not in the final time-step of the simulation after eliminatingcodemation GASOLINE those stars and which evolution (Wadsley from et al. the 2004) cooling for 10of hot Gyr. gas after the last stars do not migrate from their birth place. Dash–dotted line: profile mea- possessingan upbend) a in Type the colours II profile. at large The galaxies radii. This with upbending an essentially age profile pure in the surface density of cool gas; (3) the outerformed disk outside is the populateddisc (those with initial Galactocentric radii in excess of 57 sured in the final time-step of our simulation. Dottedgas-richWe line: profile have measured majoralso performeddoes merger, not mean simulations as that the suggested disc did not with form by inside-out.10 cosmological and In fact, 10 the par- simula- by stars that have migrated, on nearly circular25 kpc). orbits, from the exponential profile within their sample showed a plateau (and not in the final time-step of the simulation after eliminatingticlestions those per (Brookstars component, which etan‘overall’ upbend) al. thus formation 2004, in the bracketing colours of 2007). the at disc large remains ourradii. Although inside-out. This fiducial upbending However, werun. age profile in make While our use of inner disk, and consequently the break isformed associated outside the disc (those with with initial a Galactocentric radii in excess of doesdisc, thenot decrease mean that of the star disc formation did not in form the external inside-out. parts In – fact, due to the a Fig. 1.—Azimuthally averaged properties of the disk25 kpc). at 1, 5, and 10 Gyr. thesimplifications details of the gas such‘overall’decrease cooling asformationin the an volume initiallyare of the density somewhat disc remains of the spherical gas inside-out. – results resolution-depen- in However, distribution redder colours in our of mat- sharpThe top change panels inshow the the radial stellar surface mean density stellar profiles, age with profile; the dotted and lines (4) dent (Kaufmann etdisc,beyond al. the 2006), the decrease break radius. of resulting star formation in in the morphological external parts – due to a dif- ter, a lack of haloIn Fig. substructure,19, we also compare the age the profile ignoring of the galaxy with of the subsequent breakrepresenting parameters double agree exponential with fits. current The two observations. exponential components were decrease in the volume density of the gas – results in redder colours ferences,accretion we of find dark convergencebeyondone matterit would the break have and radius. if all inthe baryons, the stars modeling formed in and the disc of an – i.e. overall initial if satellites gashalo density fit simultaneously to the profile, excluding the innermost and outermost regions. cooling, star formationwereIn Fig. not and accreted.19, we its also It dependence compareis apparent the that age for profile on this of case, gas the thegalaxy surface accretion with the of den- The point of intersection of the two exponentials is taken as the break radius. profile that mimicssatellites thathas little of effect the on the dark stellar population matter, gradients. studying the ide- 2. SIMULATION METHODOLOGY sity, and disk dynamics.one it would We have therefore if all the stars formed consider in the disc the – i.e. conclusions if satellites In all panels, the vertical lines indicate the location of the break. The second alized isolated casewere not allows accreted. It usis apparent to analyze that for this case, in detail the accretion the of important and predictions of thissatellites Letter has little numerically effect on the stellar population robust. gradients. The10 6 par- Resolvedrow shows the stellar surface population density of cool data gas. In in the disk third outskirts, row, we show which the SFR are dynamical processes7.2.3 Metallicity affecting gradient the evolution of a massive iso- density calculated from the stellar mass formed in the previous 10 Myr. The ticle resolution used here represents a compromise between nowmean becoming stellar age available (normalized (e.g., by time Ferguson of output) as et a al. function 2007; of Barker radius is et lated disk. TheThe lessons metallicity gradientwe learn in the disc from and its evolution this idealized with time pro- case will adequate statistics7.2.3vide for constraints Metallicity detailed to gradient our analysis understanding in of the formation outer and disk evolu- and shown in the bottom row. al. 2007; de Jong et al. 2007), provide strong constraints on computationallater be applied costThetion for tometallicityof galaxies. 10 galaxies Gyr gradient The presenceof in evolution. theevolved of disc a metallicity and its evolution in gradient a with full in thetime Milky cosmological pro- videWay constraintsis widely accepted, to our understanding although its exact of the slope formation and shape and remain evolu- theories of break formation. Therefore, the inclusion of SF and contextOur simulation (R. Ros shouldˇkarcontentious et be (Chiappini, al., thought in Matteucci preparation). of & as Romano modeling 2001; Andrievsky disk for- formation which is caused in our simulationFigure by a 20. rapidComparison decrease of the SFR density with time in four different tion of galaxies. The presence of a metallicity gradient in the Milky feedback is required to assess break formationregions models. of the galaxy disc For – black: this between 10 andmation 15 kpc; dark and red: between evolutionWayet al. fromis 2002). widely A the accepted, related cooling issue although is the itsof evolution exact hot slope gas of thisand after shapegradient remain the with last time; this has been approached from both the theoretical (Goetz & in the surface density of cool gas; (3) the outer7Figure and disk 9 kpc; 20. Comparison purple: is populated between of the 5 and SFR 7 kpc density and orange with time between in four 3 and different 5 kpc. contentious (Chiappini, Matteucci & Romano 2001; Andrievsky reason we have run simulations of gas cooling, collapsing, and gas-rich major merger,Koeppen as 1992; suggested Koeppen 1994; by Molla, cosmological Ferrini & Diaz 1997; simula-Henry by stars that have migrated, on nearly circularregions orbits, of the galaxy from disc – black: the between 10 and 15 kpc; dark red: between3. BREAKet al. 2002). FORMATION A related issue is AND the evolution THE of OUTER this gradient DISK with forming stars inside a live dark matter halogradient7 and within 9 kpc; Another purple: which way between to appreciate 5 and the 7 kpc thisand orangetions is by betweenstudying (Brook 3 theand 5radial kpc. et al.time;& Worthey 2004, this has 1999; been2007). Chiappini, approached Although Matteucci from both & Romano the we theoretical 2001) make and (Goetz obser- use & of inner disk, and consequently the break isprofile associated of the birth parameter, with a defined as the current versus aver- Koeppenvational perspectives 1992; Koeppen (Maciel 1994; 2001; Molla, Friel Ferrini et al. & 2002; Diaz Maciel, 1997; Henry Costa gas is initially in hydrostatic equilibrium. Thisagedgradient star has Another formation the way rate further tob appreciate(SFR/ SFR thissimplifications is) (Kennicutt, by studyingIn TamblynFigure the radial & such 1, the& asWorthey Uchida an break 2003;1999; initially Chiappini, Stanghellini is already spherical Matteucci et al. 2006), evident & distributionRomano but it 2001) is as still and soon not obser- of clear mat- as a stable sharp change in the radial mean stellar age profile; and= (4) ! " Congdonprofile of 1994). the birth Instead parameter, of plotting defined the as currentter, the current SFR, a lack we versus averaged of aver- halovationalwhether substructure, perspectives the metallicity (Maciel gradient the 2001; inignoring our Friel Galaxy et al. 2002; flattens of Maciel, subsequent or steepens Costa advantagebreak parameters of making agree no with assumptions current observations. abouttheaged star the star formation formation angular over rate theb mo- last(SFR 1/ Gyr.SFR This) (Kennicutt,disk is mainly forms Tamblyn due to the & at 1&with UchidaGyr, time. 2003;Measurements moving Stanghellini using outward et H al.II 2006),regions, but as B-stars it the is still and disk notplanetary clear grows and = ! " 1 relativeCongdon few 1994). number Instead of particles of plotting with the ages currentaccretion below SFR, this value we averagedof outside dark matterwhethernebulae find the and metallicity gradients baryons, ranging gradient from and in our an Galaxy0.04 initial to flattens0.07 gas or dex steepens densitykpc− mentum distribution within the disk, which can strongly affect persisting throughout the simulation.∼− A sharp∼− drop in the local thethe break star formation radius. We over plot, the in Fig. last 21,1 Gyr. this This parameter is mainly as a function due to the of with(Afflerbach, time. Measurements Churchwell & using Werner H II 1997;regions, Gummersbach B-stars and et planetary al. 1998; profile that mimics that of the dark matter, studying the1 ide- its subsequent evolution2. SIMULATION (Debattista METHODOLOGY et al. 2006).relative few number of particles with ages belowSFR this value is alwaysoutside presentnebulae find gradients at the ranging break from radius.0.04 to The0.07 dex drop kpc− in SFR is ∼− ∼− the break radius. We plot, in Fig. 21, this parameteralized as isolated a function of caseC 2009(Afflerbach, allows The Authors. Churchwell us Journal to compilationanalyze & WernerC 1997;2009 in detail RAS,Gummersbach MNRAS the398, et important al.591–606 1998; We construct initial conditions as in Kaufmann et al. (2007). not due to a% volume density threshold% for star formation, but Resolved stellar population data in disk outskirts, which are dynamical processes affecting the evolution of a massive iso- The initial system consists of a virialized spherical NFW dark is instead associatedC 2009 The Authors. with Journal a compilation rapidC 2009 decrease RAS, MNRAS in398, the591–606 gas surface now becoming available (e.g., Ferguson et al. 2007; Barker et lated disk. The lessons% we learn from% this idealized case will matteral. 2007; halo de (Navarro Jong et et al. al. 2007), 1997) provide and an strong embedded constraints spherical on laterdensity be applied (the star to formation galaxies evolved follows in a aKennicutt-Schmidt full cosmological law at hottheories baryonic of break component formation. containing Therefore, 10% the of inclusion the total of mass SF and and contextall times) (R. Ros andˇkar a etcorresponding al., in preparation). sharp increase in the Toomre Q followingfeedback the is required same density to assess distribution, break formation which models. at the For end this of parameter. We verified that the break is not seeded by our star thereason simulation we have yields run simulations a disk mass of fraction gas cooling, of collapsing,5%. The and mass formation recipe by running several simulations with different 12 ∼ 3. BREAK FORMATION AND THE OUTER DISK withinforming the virialstars inside radius a is live10 darkM, matter. A temperature halo within which gradient the in values of the threshold density and found that the location of thegas gas is component initially in hydrostatic ensures an equilibrium. initial gas pressure This has theequilibrium further theIn Figure break 1, did the not break depend is already on the evident particular as soon value as a stable used. Since foradvantage an adiabatic of making equation no of assumptions state. Velocities about the of angular gas particles mo- diskdensity forms is at inversely 1 Gyr, moving proportional outward as to the radius, disk the grows cooling and time arementum initialized distribution according within to a cosmologicallythe disk, which can motivated strongly specific affect persistingincreases throughout outward. the By simulation. construction, A sharp the drop angular in the momentum local is angularits subsequent momentum evolution distribution (Debattista with etj al.r 2006). and an overall spin SFRdirectly is always proportional present at the to cylindricalbreak radius. radius, The drop which in SFR means is that parameterWe constructl p ( j/ initialG)(FE conditionsF/M 3) 1/2 p as0.039 in Kaufmann∝ (Bullock et et al. al. (2007). 2001). nothigher due to angular a volume momentum density threshold material for will star formation,take longer but to cool. The initial system consists of a virialized spherical NFW dark is instead associated with a rapid decrease in the gas surface matter halo (Navarro et al. 1997) and an embedded spherical density (the star formation follows a Kennicutt-Schmidt law at hot baryonic component containing 10% of the total mass and all times) and a corresponding sharp increase in the Toomre Q following the same density distribution, which at the end of parameter. We verified that the break is not seeded by our star the simulation yields a disk mass fraction of 5%. The mass formation recipe by running several simulations with different 12 ∼ within the virial radius is 10 M,. A temperature gradient in values of the threshold density and found that the location of the gas component ensures an initial gas pressure equilibrium the break did not depend on the particular value used. Since for an adiabatic equation of state. Velocities of gas particles density is inversely proportional to radius, the cooling time are initialized according to a cosmologically motivated specific increases outward. By construction, the angular momentum is angular momentum distribution withj r and an overall spin directly proportional to cylindrical radius, which means that parameterl p ( j/G)(FEF/M 3) 1/2 p 0.039∝ (Bullock et al. 2001). higher angular momentum material will take longer to cool. [email protected]

Vertical vel. disp. vs. Metallicity

250 NOTE: put ALL PLOTS YOU HAVE HERE!! Sa Sab Sb 200 Sbc

] Sc

-1 Scd

150 [ km s z !

100

50

250 0 Sa -1.5 -1.0 -0.5 0.0 0.5 Sab [MgFe][Fe/H] Sb talk on: www.iac.es/project/traces Sbc 200 Sc Scd R

! 150

100

50 0 5 10 15 20 age [Gyr] [email protected] Ages and metallicities via line strengths sigma drop region bulge disk

AGE

Z

talk on: www.iac.es/project/traces [email protected]

The idea

Unravelling the nature of bars & bulges: observing secular evolution in action

Use integral-field spectroscopic observations to study kinematics and stellar populations in two dimensions of (mainly) barred galaxies Reveal the mass distribution and star formation history of the chosen galaxies Link the 2D stellar dynamics and stellar populations to constrain scenarios for the secular evolution of galaxies under the influence of bars

talk on: www.iac.es/project/traces [email protected]

Observations performed

2 successful observation proposals and observation at the WHT

Data obtained and preliminarily reduced after the run: March 2012 for 3 out of 5 galaxies observed, binned to S/N = 40; SAURON mosaic on top of the photometric image from SDSS

talk on: www.iac.es/project/traces [email protected]

Outlook

Unravelling evolution processes by comparing with simulations

From simulations: bars evolve and become stronger with time leaving an imprint in the LOSVD Growth as a consequence of angular momentum transfer between the different components of the galaxy: disk, bar, dark matter halo. Different stages in the time evolution of an early-type barred galaxy in one of our numerical simulations (Martinez-Valpuesta et al. 2006) :

talk on: www.iac.es/project/traces [email protected]

Control sample and disk heating focus Disk-heating (Shapiro & Gerssen 2003 and 2012)

NGC 2280 (Scd) NTT NGC 3810 (Sc) NGC 4030 (Sbc) NGC 1068 (Sb) KPNO NGC 2775 (Sa/Sab) NGC 2460 (Sa) pPXF (Cappellari & Emsellem, 2004) Gandalf (Sarzi et al. 2006) Mean ages and Z derived with Starlight (Roberto Cid Fernandes, 2007)

talk on: www.iac.es/project/traces

age [Gyr] Z [email protected]

Outline

Introduction

Sample

Methods

Preliminary Results and Discussion

Outlook

talk on: www.iac.es/project/traces [email protected]

talk on: www.iac.es/project/traces