Goals of SummaryOverview of points from talks Some extra itemsIdentify some problems for discussion

Secular Evolution in Disk GalaxiesSummary: Ken Freeman

The main themes:Formation of bulges by secular and non-secular processesSecular processes : rings, orbits, inflowStellar population issues

Ringberg Workshop May 2004

The expression "secular evolution" .... what does it mean ?

I think it is intended to mean "dynamical evolution that is slower than the

dynamical timescale of the disk"

I think it was used here by many to mean "dynamical evolution that occurs after the disk has formed"

so I will use it in this sense

but

Appreciation of secular evolution in galaxies has grown over last ~ 20 years

Its significance for properties of barred galaxies was understood early

Accepting importance of secular processes for bulge formation (including disk instabilities) is more recent - not complete yet

JK&RK ARAA, and this workshop, will help a lot

Kormendy overview (based on ARAA article)

Secular evolution and the growth of pseudobulges.

• SE now becoming more important than hierarchical effects

• bars grow by outward transfer of J - orbits elongate, pattern speed drops. Gas at low R goes in to center, gas at intermediate R goes to inner ring, gas at large R goes to outer ring near OLR

• dustlanes along bars identified as shocks - lead to energy loss by gas and infall to center

• inner outer and nuclear rings usually star-forming

• circumnuclear star-forming rings should generate pseudobulges on 1-3 Gyr timescales.

• pseudobulges as systems above the oblate rotator curve in the (V/ - ) plane.

• exponential bulges

• bar destruction by growth of central mass. lenses as defunct bars

Q: how do the smaller disky ellipticals fit into this picture ?

Athanassoula : N-body view of secular evolution

J-exchange drives bar growth: J goes into outer diskand halo near resonances, or to interacting galaxy

The live halo is an important element in the growth and evolution of the bar.

This realisation has really changed perception of role of halo in dynamics of barred galaxies, from suppressing

bar growth (rigid halo) to enhancing bar growth (live halo)

N-body boxy bulges are bars(or at least the inner ~60% of the bars)

Side-on bars are strongly B/P whileend-on bars appear as relatively round bulges.

Rotation is cylindrical in simulations, as in reality

These processes are a mixture of secular and rapid : bar growth is slow,

bending instability is fast, subsequent settling of the bar/bulge slow again.

Patsis : boxy isophotes in barred galaxies

Periodic orbit properties define the structure of products of dynamical evolution

See some of the fine details of 3D periodic orbits reflected instructure of unsharp-masked images of real peanut galaxiesaway from the plane

Detailed structure of face-on SB's (eg boxy isophotes ofbar) relate to morphological properties of orbits

Debattista

Simulations of the buckling instability for anisotropic bar and development of peanut bulge.

Bar survives the process. Buckling driven by anisotropy, heats the disk vertically

to reduce the anisotropy

h4 provides potential diagnostic of face-on peanut, even in axisymmetric systems.

Possible relation of buckling instability to observed double exponential disks, often described as truncation, seen near OLR.. Not 1-1 : see truncations without bar and bars without truncation

Example of truncation in M33 : origin of truncation is significant problem - is it an effect of secular evolution ?

The truncation of

M33's disk

(Ferguson et al 2003)

M33 is a puredisk galaxy in

the Local Group

M33 Surface Brightness Profile:

• i-band surface photometry out to R ~ 35'

• profile extended to R ~ 60' using star counts

Disk Truncation

V~31 mag arcsec -2

surely the deepest surface brightnessprofile ever measured for a pure

disk galaxy

Ferguson 2003

sharp decrease in surface brightness beyond 5 scalelengths..

Bureau

Observational and N-body overview of bar-drivenevolution in B/P bulges : ~40-50% of bulges are B/P

Structure of V, h3 gives a strong kinematic signature of end-on bar where photometric signatures are absent

Good correspondence between observed stellar kinematicsin B/P bulges and the simulations.

These bars are not flat - by definition. See changes in the vertical scale length with radius.Often have thick bar with embedded central disk. Unlike the Ferrars bars widely used for orbit studies

The typical B/P bulge structure has bow tie structure, and includes a flat sort of pseudobulge with young population

Some evidence in NGC4526 for vertical extension ofyoung stars - maybe resonant heating near bar ends

Q: Is there such a thing as a flat bar ? How important is resonant heating ?

Buta : morphology of barred galaxies

Rings usually blue - enhanced star formation relative to backgroundNuclear rings have range of shapeSecondary bars have no preferred orientation

Production of rings from sticky particle codes - predict that SB(s) galaxies have weaker bars.Observational indication that bar is actually stronger for the SB(s) galaxies

Examples of galaxies that might have slow, medium and fast pattern speeds (give various numbers and locations of resonances)slow pattern speed <=> nuclear rings, medium pattern speed <=> inner ring but no nuclear ring.

Underlying the detailed morphology are fundamentalproperties like • the presence of secular evolution• the presence of a live halo

In systems with slow pattern speeds, we could expect up to 5 resonant features - maximum RB has seen so far in any one galaxy is 4 !

Sparke - double and multiply-barred galaxies

~ 25% of 38 SB0, SBa galaxies are double/multiply barredSecondary bars are randomly orientated - not bluer thansurroundings, so probably longlived

Look for orbits that reinforce the secondary bar in presenceof a time-periodic potential - eg in ordinary bars, x1 familyextends to CR and provides backbone of bar - what is analog for secondary bars ?

Developed concept of stable invariant loops - require smallbar to be inside the ILR of the outer bar and well inside itsown CR

Regan - nuclear ring formation in SB galaxies

Hydro simulations of flow in fixed potential. The nuclearring is feature trapped between the ILRs, growing viainflow associated with the near-radial shocks. Rings form best when galaxy has a central mass concentration and bar is not too strong.

Nuclear rings are associated with x2 orbits which extend overregion between IILR and OILR and are needed for offset shocks. Rings start to form at the outer x2 orbit and gradually evolve inwards as lower-J gas is accreted - associated inflow rate up to ~ 0.5 M_sun yr -1. Inflow requires presence of nuclear rings

Gas prefers to associate with the more circular x2 orbit but do see rarer x1 rings at location of largest non-looped x1 orbit..

Balcells - nuclear & global properties and secular evolution

HST NIR images for 19 SO-Sbc unbarred galaxies, SAURONspectra, groundbased optical images

Sersic fits: ~ r1/n : finds very few n=4 bulges. Mean isn=1.7 plus point source (plus inner exp). Point source islike 10-20 globular clusters. Low Sersic n => not much mergergrowth for these bulges

30% have extended nuclear cpt with (0) as bright asthe Sersic component - inner bar/disk

Well defined scaling laws for most parameters with Lbulge,but not much with type between SO and Sbc

Strong scaling laws with Lbulge => disk grew around existingbulge, so bulges in place 10 Gyr ago

Q: how would the M31 bulge look in this analysis ?Unambiguous evidence for r1/4 law - would this kind of analysis have found it ?

Pritchet & van den Bergh 1994

M31 surface brightnessdistribution

Carollo

Massive bulges have z > 1/2 zsun, ages 8-10 Gyr Smaller bulges have ages 2-5 Gyr.ie, many small bulges are younger than the more massivespheroids, with ages like surrounding disk - see age spread

Deep HST images show presence of normal disk galaxies atz=1, with disks and bulges in place, scale lengths notmuch evolved since. But bulges at z~0.5 are systematically bluer than ellipticals

Bars are now abundant up to z~1, despite earlier claimsthat bars are present only for z < 0.7.

For z = 0.5 to 1, ( bulge age - disk age) < 2 Gyr forintermediate to later type disks. Conclude that bulges in Sb and later galaxies are products of internal disk evolution

Bulge simulationsIn fixed halo potential, bar buckling produces bulgeswith V/ both above and below oblate rotator line.In live halo plus SPH, outcome depends on gas fraction.• Gas ~ 10% gives rounder smaller flatter bulges• Gas ~ 50% with radiative cooling gives very clumpy system with strong spiral structure (not bar). Very fast evolution, generates an old bulge from disk.

Erwin - bars and secular evolution

Bar destruction via central mass growth - might expect process like SBc -> SABb -> SAa ie fewer bars at earlier types. However, bar fraction isobserved not to vary much with Hubble type. Littleroom for bar suicide.

RC3 survey: SO-Sb bars have mean semilength 1.4h Sc-Sd mean semilength is 0.6hunlikely to be evolutionary phenomenon

Sample of 14 SB0 galaxies - about half lie above theoblate rotator line.

Sample of disks, about half with type I profiles - mostshow no truncation out to > 5h

Falcon-Barroso - SAURON results

See the complexity of inner regions : a range of central features - kinematic twists, decoupled cores, counter-rotatingdisks, pseudobulges [flat (r], non-axisymmetric objects inmore than 50% of examples, central stellar disks in 75%

N7332 - boxy edge-on bulge. Shows cylindrical rotation,KDC, hint of central disk. Gas is irregular. Stellarage homogeneous over the bar - near-solar, age 3-5 Gyr.

Minor axis slit spectra of 19 galaxies. Some show peakedsigma(r) - classical bulges. Others have flat sigma(r) -dominated by rotation, probably pseudobulges.

Barden - ACS survey of 30 x 30 arcmin field

Make Sersic fits - take galaxies with n < 2.5 to be disks

See mean surface brightness V increase by ~ 2 mag from z = 0.2 to z >1 - good agreement with expectation from theory. Much as expected from passive evolution.Little change in surface density with z, suggests that galaxiesincrease in size as they accrete gas - inside out disk formation

This was a report on how disks look at different epochs -some discussion about whether this is evolutionary

Binney - secular evolution of galactic disks

Heating effect of transient spiral waves: stars exchangeangular momentum, mainly across CR - gives muchradial migration but with little associated heating. Theheating that does occur takes place near ILR. GMCs don'theat but redistribute spiral arm heating from plane to z

Large spread in [Fe/H] near the sun is evidence that this radial migration does occur

The spiral structure generates structure in UV planefor Hipparcos stars - some of this structure is bar-driven

Q: Do we now undertand the mechanism for disk heating. Disk heating is an important secular effect.

See that significant heating takes place near the sun for

stars with ages between 1 and 2 Gyr. •

Binney-Sellwood heating requires the transient spiral arms to have ILR near the solar radius - is this plausible ?

Freeman 1991; Edvardsson et al 1993; Quillen & Garnett 2000

Velocity dispersionsof nearby F stars

old disk

thickdisk

important

Disk heating saturates at 2-3 Gyr

Gerhard - clump and bulge formation in gas rich galaxies

Bulge formation options - mergers, secular evolution, clump instability in gas-rich disk. Note how clumpy somegalaxies are at z ~ 1.5.

Models include gas and stars with interaction network, plusdark matter. Generates a few clumps per galaxy - dynamical friction and spiral arm torques funnel clumps to center to form a bulge. Fast bulge formation, as observed by Carollo. About 10% ofbaryons end up in bulge. Can add more by secular processes.

Range of metallicities for clumps is -0.7 to -1.Some stars have [Fe/H] > 0, [Mg/Fe] > 0

To produce pure disk galaxies, need to suppress clumpformation - hot gas, low cooling rate

Fragmentation efficiency depends on cooling rate - fastercooling gives more clumps. Models with lower cooling rate(hotter gas) give smooth distribution of star formation - no clumps - till disk goes unstable.

Moore - harassment

Environmental effects can speed up secular evolution

Showed convincing model of Magellanic Stream :hydro-tidal phenomenon as LMC orbits around MW

Effect of the cluster potential - smooth and lumpy components - on evolution of disk galaxies. Note the spiral structure seen in unsharp masked images of some low-luminosity ellipticals. Many stars are lost from disks - bending, bar instability, leads to round objects. Large variety in V/ - distribution.

Rosolovsky - atomic -> molecular transition in LG galaxies

In M33, MCO/Mvir for GMCs is independent of [O/H]

GMC populations in MW, LMC, M31, M33 look asif they come from the same population

Mass function for M33 GMCs is steeper in galaxies withhigher Toomre Q.

Conversion of gas into GMCs may be main issue in star formation

GMCs in M33 lie on peaks of the HI distribution

GMCs are significant in secular evolution of galaxies

Renzini : Stellar populations in the bulges of MW and M31

Zoccali MW bulge field at b = -6o : Old stellar population, like bulge globular clusters. No evidence for intermediate-age stars - the AGB star density is as expected for metal rich globular clusters

Q: the bulge stars are very old, but does this mean non-secular formation of the bulge, before the disk ? Is the bulge structure also so old ? How can we tell ?

M31 bulge has similar K-band LF to the MW - also old

Example of some bulges at z ~ 1.6 that are already 1 Gyr old, so formation z ~ 2.5

A few slides on

the structure of the Milky Way bulge

The Galactic Bar- Bulge

small exponentialbulge - typical of later-type galaxies.

Unlike the large r1/4 bulge of M31

Launhardt 2002 Pritchet & van den Bergh 1994

M31

The galactic bulge is rotating, like most other bulges: (Kuijken & Rich (2002) HST proper motions)

Beaulieu et al 2000K giants from several sourcesand planetary nebulae (+)

Velocity dispersion of bulge andinner disk are fairly similar- not easy to separate inner diskand bulge kinematically

Bulge ends at |l| ~ 10o

Inhomogeneous collection of photometric ( ) and spectroscopic ( ) mean abundances - evidence for abundance gradient along minor axis of the bulge

Minniti et al 1995

( kpc )

Abundance gradient inthe bulge

Zoccali et al (2002)

Near the center of the bar/bulge is a younger population,

on scale of about 100 pc : the nuclear stellar disk (M ~ 1.5 x 109 M_sun)

and nuclear stellar cluster (~ 2 x 107 M_sun )in central ~ 30 pc. (Launhardt et al 2002)

~ 70% of the luminosity comesfrom young main sequence stars.

Maraston, Thomas : Stellar population propertiesof spheroids, including the MW bulge

Enhanced [Mg/Fe] ~ +0.3 needed to model the Fe-Mgdistribution in spheroids => short star formation timescale,~ 1 Gyr.

[Mg/Fe] correlates with velocity dispersion : massivesystems form most rapidly

Need AGB stars to synthesize intermediate-age populations. No evidence for intermediate agepopulation from M31 bulge SED

Bulges and ellipticals are similar in their stellar populations.They follow an age - (velocity dispersion) relationYoung apparent age at low probably means rejuvenation : underlying old population is present also.No trend of properties with Hubble type, just with : ie mass drives the stellar properties, not T

Q: is there a definitive stellar population test of concept that bulges form from disk instabilities ? Are chemical and age gradients the key ?

Peletier : Secular evolution and stellar populations

Color maps do not show the peanut structure : the bulges and inner disks appear very similar. Early-type bulges are old (~ 10 Gyr) with small scatter (~ 2 Gyr).

Smallest bulges lie slightly off the fundamentalplane for the old systems.

Bulges of S0-Sb galaxies are like ellipticals in mostproperties (at similar luminosity), except for smaller Sersic index (1-2.5 vs 4).

Later-type bulges are different : star formation in rings,shallow surface brightness profiles ... (cf Carollo surveys)

Courteau/MacArthur : Evidence for secular evolution in non-barred galaxies

Exponential bulges in later-type spirals showstrong correlation of re,bulge with hdisk

Mean value of re,bulge / hdisk is 0.22, independent ofwavelength and Hubble type, consistent with secular evolution models

Brodie : Constraints from star clusters on secular evolution in spirals and lenticulars

Tight link between the red mode of cluster formation andformation of the bulge - related to our problem in some way, though cluster formation is still poorly understood

Faint fuzzies : old loosely bound clusters seen in annularregion of some SB0 galaxies. Apparently longlived, despiteapparent fragility. Survival suggests association with low-e orbits. Formation in resonance rings would indicate presence of bar at the time of formation, z > 2.

Bosma, D'Onghia, Burkert

Bulgeless disks - the other side of secular evolution

Real problems understanding ...(i) how such disks can form (the angular momentum problem) (ii) how they survive without going unstable and forming bulges - role of the dark halo, anisotropy limits ...

Why do we see bulgeless galaxies at all ?

Exciting field - serious problems - much work to do in observations, dynamical theory, galaxy formation theory ...

Two secular processes that

did not receive much attention ...

1. Effect of the potential of a slowly rotating triaxial dark halo : these are common in LCDM simulations

2. Nucleation in spirals

Secular effects of rotating triaxial dark halo - ubiquitousin LCDM simulations (Bekki & Freeman 2003)

eg spiral structure in far outer regions of gas-rich diskgalaxies, where Q, X too large for spiral structure

NGC 2915 - HIout to 22 scalelengths-see spiral structureBureau et al (1999)

Another example of far-outer spiral structureNGC 6946 - Oosterloo et alvery deep WSRT HI image

andanotherexample

NGC 5055

Oosterlooet al

Nuclei mostly very central - why ?

Nuclei of spirals like M33 and NGC 7793 (Walcher) show extended history of star formation. Looks like a secular evolution process.

What is cause of nucleation ?

Nucleation in spirals - probably secular process

ngc1493

Nuclei of late-type spirals

Böker et al (2002) survey: 59/77 late-type spirals (T > Sc)have compact nuclei very close to center, as compact and massive as globular clusters.

Most of these late-type spirals have no visible bulge.

Nuclear clusters are mostly isolated - no spiral arms, dust lanes or other indications of a dynamical center.

Böker et al 2002

Offset of nuclei from center of disk isophotes

The nuclei lie very close to the galactic

centers

Several authors make a point about the central locationof these nuclei - how do they know where the center is,in the shallow central potential well of an exponential disk.

Exponential disk is not as shallow as some ! d/dr does not as rRotation curve slope dV/dr is singular at center

Density M(r) ~ V(r) ~ d/dr ~

Keplerian const r -1/2 r -2

(r) ~ r -1 r const r -1

(r) ~ e-r/h r 2 r 1/2 const

= const r 3 r r

The nucleus of the late-type spiral NGC 7793

Black is the observed UVES spectrum: R = 32,000Red is a single-burst population, age 108 yr - not very good fitBlue is a mixed-age fit - good fit

Dynamical M/L is 2.2

Single burst M/L is 0.5

Mixed pop M/L is 2.8

Walcher et al 2003

This shows the fractional

mass and

luminosity in the mixed components

The young component contributes 1.7% of the mass and its age is0.8% of the age of the universe. Suggests that star formation in this nucleus (NGC 7793) is an ongoing (secular) event, as in M33

Simon White's Questions & Issues

Are bulges built before or after disks ?

How do galaxies at intermediate redshift (as seen several Gyr ago) map into today's galaxies - do we see the bulge-forming events ?

How much of the light in the local universe comes frombulges ?

Issues

Rundown of merging vs exhaustion of gas

Secular stellar dynamics vs starformation/gas driven evolutionRole of accretion/inflow in secular evolution

BH-bulge/pseudobulge relation and BH feeding.

Q: how do the smaller disky ellipticals (so similar to bulges in many properties) fit into the secular evolution picture ?

Q: Is there such a thing as a flat bar ? How important is resonant heating ?

Q: What would a Sersic fit (including a central point source) give for the bulge of M31 ? M31 has an unambiguous r1/4 bulge - would this kind of analysis have found it ?

Q: Disk heating is an important secular effect. Do we understand the mechanism for disk heating in the solar neighborhood ?

Questions arising

Q: is there a definitive stellar population test of concept that bulges form from disk instabilities ? Are chemical and age gradients the key ?

Q: the MW bulge stars are very old, but does this imply non-secular formation of the bulge, before the disk ? Is the bulge structure also so old ? How can we tell ?

Two related questions ...

Thanks to Andi, John & Ralf for

a great workshop