Ana səhifə

Contents §0 Introduction: Importance of Galactic Bars §1 General Observed Properties of Barred Galaxies §2 Dynamical Effects of Bars


Yüklə 231.7 Kb.
səhifə3/3
tarix09.05.2016
ölçüsü231.7 Kb.
1   2   3
§6 High-z Bars: Cosmological Perspectives
fundamental questions: When and How did barred galaxies form?

 Observations of distant galaxies (with large look-back times)


6.1 Observations
◆ Abraham et al.(1999, MN, 308, 569)

46 WFPC2-imaged galaxies (in HDF) in F814W (I-band)

rest-frame optical bar fraction decreases from 24% at z ~ 0.2-0.7 to 5% at z >0.7
Estimate of Bar Strength (using two isophotes)

1) pixels with > 1.5 σ above sky level are used.

2) 1% and 85% of the maximum flux level used to define outer and inner regions.

3) calculate second order moments

4) calculate (b/a)outer and (b/a)inner and PA difference φ.

5) correction for inclination i = cos-1[(b/a)outer ] →(b/a)2


1)Test for z ~ 0

Frei et al. Sample (1996) (Fig.1)

For inclination < 60°‘(b/a)2=0.5’ effectively distinguish between barred and unbarred galaxies.


  1. Test for HDF condition

- Frei et al. Sample redshifted artificially to 02 can be determined accurately. →I814W = 23.2

- Bright HDF galaxies degraded to the magnitude limit (Fig.2)







Analysis of HDF Sample:

I band images

z <1 to avoid band shift effect and retain high S/N

Spiral galaxies are extracted using A and C (see Appendix)(and i < 60°)

→ 20 HDF-N and 26 HDF-S spirals

→ 9 HDF-N and 13 HDF-S spirals are strongly barred (48%)


[Caveat: Barred galaxies may have been preferentially excluded in extracting spiral galaxies using A and C]

 All galaxies in HDF-S/N to I814W = 23.2 checked.

→ Visual classification and A-C spiral selectin are consistent.
Results:

Fig.4  Bar fraction decreases as z increases



[Caveat: Unbarred galaxies are intrinsically more luminous than barred

galaxies so that high-z sample is dominated by unbarred galaxies]

→ Local (z ~ 0) galaxy catalogues (e.g., RC3, Revised Shapley-Ames Catalog) do not support this.

 SBc are slightly less luminous than SAc (van den Bergh 1998, conference proceedings)

Interpretations:

1) High-z disks are not massive enough to be bar-unstable.

2) High-z disks have large random motions and cannot create bars.

3) Bars are more efficiently destroyed at higher z (e.g., by galaxy mergers, especially minor mergers which turn barred spirals into unbarred spirals)

4) Buckling Instability or Central Mass Growth turns bars into bulges so that

we see few bars at redshift proved in this study.

[3), 4) need second bar generation to explain high bar incidence at z ~ 0]


Appendix: C & A (Abraham et al., 1994, ApJ, 432, 75)
A = area enclosed by the isophote at a given (typically 2σ) level above the sky

Mxx ≡ /, Mxy ≡ /

Myy ≡ /
Define `normalized radius' by

r2 = Mxx x2 + Myy y2 - 2 Mxy xy , so that E(1)=A.


Define `concentration parameter C' by

C ≡ /


where α is the 'shrinking factor'(usually α=0.3).
Define `asymmetry parameter A' (Abraham et al. 1996, ApJS, 107, 1)

Rotate Iij by 180° about center  Jij

A ≡ /
A-C diagram (Abraham et al. 1996, MN, 279, L47) can separate galaxy morphology

(Fig.1, visual morphology by Richard Ellis)




◆ Sheth et al. (2003, ApJ, 592, L13)}

HDFN NICMOS (H-band 1.6 μm) to examine effects of

1) Bandshifting

2) Limited spatial resolution


Bandshifting

NICMOS data provide rest-frame V through I-band images for z > 0.7.


Important because

1) Bars are composed of old populations

2) Bars often have non-uniform dust obscuration and peculiar star formation

(→ Fig.1)



Analysis:

1) 904 galaxies with photometric or spectroscopic redshift in HDFN

(206 at z<0.7, 226 at 0.7

2) 136 candidate disk galaxies selected using V, I, H images

(41 at z<0.7, 95 at 0.7

3) IRAF ELLIPSE routine applied to identify bars

(= constant position angle and increasing ellipticity)

Bar will be missed if

- the galaxy is highly inclined

- bar position angle is the same as the disk

- underlying disk is too faint

- data have poor resolution

 Fraction of barred galaxies will not be overestimated


Results:

- For z<0.7: 5 barred spirals + 2 candidates

 consistent with Abraham et al. (1999) (7 barred spirals for z<0.7)

- For z>0.7: 4 barred spirals (Fig. 2, Table 1) + 5 candidates

 Abraham et al. (1999) 2 barred spirals for z>0.5

 More barred galaxies despite more conservative bar identification than

Abraham et al.(‘ellipse fitting over entire image’ vs ‘only two isophote’)
Introduce cutoff at I(AB)=23.7 (same as Abraham et al. )

 For z>0.7: 3 barred galaxies among 31 disk galaxies

 consistent with bar fraction not decreasing at z>0.7
Spatial Resolution and Visibility of Bars 

Typical bar size in nearby galaxies ~5 kpc

NICMOS bars are larger than 12 kpc

Bar fraction (> 12 kpc):



- 4/95 for z>0.7

- 3/31 for z>0.7 (for I(AB)<23.7 as in Abraham et al.)

higher than the local value 1/44 (SONG)

[SONG = Survey Of Nearby Galaxies (Regan et al. 2001, ApJ, 561, 218)

44 Sa-Sd galaxies brighter than 11.0 B-mag, with V < 2000 kms-1

and i < 70°(originally selected for BIMA CO survey)]
Implications:

1) Massive disks were already present 7 Gyr ago.

 consistent with star formation history

2) Interactions induced bars


◆ Elmegreen, Elmegreen & Hirst (2004, ApJ, 612, 191)

ACS F814W imaging of Tad-pole galaxy field (3.9' × 4.2')

186 galaxies with major axis larger than 0.5”(= 10 pixels)
bar detection :

1) visual detection + (a maximum ellipticity accompanied by a change in

position angle at the same radius) → 'clearly barred'

2) twist of isophote in center → 'inner isophotal twist'


dependence of bar fraction on inclination and size taken into account

(i.e., bar fraction decreases in highly inclined or small galaxies)


Results: constant bar fraction (20-40%) at z ~ 0 - 1.1


◆ Jogee et al. (2004, ApJ, 615, L105)
Abstract

1) 'How bars evolve over cosmological times' has yet to be addressed.

- a recent phenomenon or abundant at early cosmological epochs?

- short-lived or long-lived?

- recurrent (dissolve and re-form) or one-time event?

- stellar bars ⇔ hierarchical clustering of galaxy evolution, underlying disk evolution ?

2) investigate frequency of bars out to z ~ 1 using 1590 galaxies from GEMS.

- two color images from HST ACS (Advanced Camera for Surveys)

- redshifts from COMBO-17

3) results:

constant bar fraction (~ 30%) at lookback time 2-6 Gyr (z ~0.2-0.7), 6-8 Gyr (z ~ 0.7-1.0), and present.

4) implication:

- cold gravitationally unstable galactic disks already present at z ~ 1

- bars have a long lifetime


Motivation: limitations of previous studies

1) Abraham et al. (1999, MN, 308, 569) suffers from

- bandshift effects (F814W observe rest-frame ~ 500 nm)

- low resolution of WFPC2 (~ 0.1” in F814W → miss <5kpc bars )

2) Sheth et al (2003, ApJ, 592, L13 =S03)

NICMOS images of 95 galaxies in HDF in F160W (H-band)

- bandshift effects much reduced

- even lower resolution (~ 0.2-0.3” in F160W) than WFPC2

3) Elmegreen, Elmegreen & Hirst (2004, ApJ, 612, 191)

ACS F814W imaging of Tad-pole galaxy field (3.9' × 4.2')

- large (0.1-0.4) error of photometric redshifts (Benitez et al. 2004, ApJS, 150,1)

F475W, F606W, F814W → photo-z

(present study: improved by using COMBO-17 high-accuracy photo-z)

- small sample size precludes absolute magnitude completeness at different z (overcome by GEMS galaxies)

- bandshift effects (minimized by two passbands)
Observations, Sample, Methodology

GEMS (Rix et al. 2004, ApJS, 152, 163)

・ two-color (F606W, F850LP) HST ACS imaging survey

・ 28' × 28' field centered on Chandra Deep Field South

・ high resolution ~ 0.05” (360pc at z ~ 0.7)


・~ 8300 galaxies at z ~ 0.2 - 1.1

・ redshift & SED available from COMBO-17 (Wolf et al. 2004)


This paper: ~ 25% of GEMS (14' × 14' ~ 30 ×HDF )

 1590 galaxies at z ~ 0.2 - 1.0 and RVega <24


Method: ellipse fitting

・ 90% of 1590 (1430) galaxies successfully fitted.

・ 10% failure = disturbed systems or low surface brightness objects

・ inclined system (i > 60°) excluded.


Selecting disk galaxies: (大規模サンプルで銀河の形態分類をどうするか?)
(1) Sersic n < 2.5 (Bell et al. 2004, ApJ, 600, L11)

GEMS F850LP images of 1492 0.65 < z < 0.75 galaxies (~ rest-frame V-band)

morphology classification:

1) by eye → E/S0, Sa, Sb-Sdm, Peculiar/Strong Interaction, Irregular/Weak Interaction, unclassifiable

2) automated galaxy classification using single Sersic model

 n = 2.5 separates early (E/S0) and late (Sb-Sdm) galaxies

(with less than 25% contamination)

('n<2.5-criterion' also tested by using artificial galaxies)


(2) Concentration index C < 3.4 (Conselice et al., 2000, ApJ, 529, 886)

113 nearby galaxies from Frei et al (1996, AJ, 111,1) catalog

aperture photometry (using circular aperture)

C ≡ 5 log [r(80%)/r(20%)] (r1/4 law →C = 5.2, exponential law →C = 2.7)

 C vs RC3 morphology
(3) rest-frame U-V < 0.8

- Coleman et al. (1980, ApJS, 43, 393)

SED of local ~ 10 spirals

- Bell et al. (2004) study of GEMS z ~ 0.7

 galaxies visually classified Sa-Sm galaxies have U-V < 0.8
Bar detection: (bar の客観的定義が必要)

(1) ellipticity (e) rises to global maximum emax > 0.25,

while PA has a plateau (within ± 20° )

(2) Beyond bar end, e must drop by >0.1, while PA changes by >10°

bar fraction fopt ≡ Nbar/Nsp-disk
Results :

Fig2


- most bars identified have e>0.4, a=0.15”-2.2” (a=1.2-13kpc)

- e<0.3 bars difficult to identify

 hereafter we discuss only e>0.4 bars

Table 1


- bar fraction constant at ~ 30% ± 6%

- no incompleteness effect

 similar bar fractions for

<-18.5 (to match OSU sample with absolute mag. range of -18.5 ~ -23.0)

and


<-19.5 (complete out to z ~ 1.0)

- bar fraction slightly higher in rest-frame I-band than B-band


Redshift dependence

1) agree with Elmegreen et al. (2004, ApJ, 612, 191) about Tad-pole galaxy field

2) do not agree with Abraham et al.(1999, MN, 308, 569)

‘dramatic decline in bar fraction’

reasons:

- small bars (a < 0.5”) missed by wider PSF in WFPC2 and NICMOS

- low number statistics

- cosmic variance

- methodology
Comparison with local Universe:

OSU sample (Ohio State University Bright Spiral Galaxy Survey)

(Eskridge et al. 2002)

- fopt ~ 37% based on RC3 'SB' bar class}

- fopt ~ 33% according to B-band ellipse fit (e > 0.4)}
Conclusion:

strong bar (e > 0.4) fraction remains similar at 30 - 37% for

- z ~ 0.7-1.0 (lookback time ~ 6-8 Gyr)

- z ~ 0.2-0.7 (lookback time ~ 2-6 Gyr)

- present day


Implications:

(1) dynamically cold disks already in place by z ~ 1 (バー不安定でバー

ができるための条件)

(2) triaxial and centrally-concentrated DM halos (predicted by CDM

cosmology) not prevalent at z ~ 1

triaxial and centrally-concentrated halo destroy bar

 (El-Zant & Shlosman, 2002, ApJ, 577, 626)

(3) large-scale stellar bars are long-lived

(4) bar ellipticity and length can evolve
Appendix:

COMBO-17 (=Classifying Objects by Medium-Band Observations in 17 Filters)

(Wolf et al. 2003, A&A, 401, 73; Wolf et al. 2004, A&A, 421, 913)

・ deep surveys of 4 field (34' × 33' each)

・ 17 passbands : 350 nm < λ < 930 nm to obtain SED and photo-z

・ telescope: MPG/ESO 2.2m-telescope on La Silla, Chile

・ complete to mR ~ 24

・ redshift accuracy:

δz/(1+z) ~ 0.02 for R ~ 22

δz/(1+z) ~ 0.1 for R > 24


6.2 Theory
Cosmological N-body simulations

Initial density perturbations  formation of dark matter haloes with gas 

hierarchical merging of halos  galaxies of various mass and size form

Disk galaxies do not form! (angular momentum catastrophe)
Cosmologically motivated numerical simulations for barred galaxies
◆Bournaud & Combes (2002, AA, 392, 83)

Effects of gas accretion and repeated bar instability


Initial disk: stars + gas particles (Q=1.5)

Accretion: increases number of gas particles


2D models: halo and bulge = fixed gravitational potential

parameters Mb/Mdisk,i Mh/Mdisk,i










Gas accretion

- Velocity = circular velocity at the outer disk + random motion with Q=1.5

- Position = outer disk edge (azimuthally uniform)

(outer disk edge moves as the accretion proceeds)

- Rate = double the disk mass in 7-10 Gyr

Halo mass partly reduced (e.g. Mh,dot= -1/3 Macc,dot)


Results:

- Generally once formed bar disappear as a result of gas accumulation at the center in absence of gas accretion (Fig.1)

- gas accretion help maintain a bar or

- gas accretion induces reappearance of bars (Fig.1,Fig.4)

depending upon halo/bulge/disk mass ratios (Fig.3)

- initial gas/star ratio in disk does not affect evolution much , and effect of

gas accretion is more important.

[Fig.4: Fourier analysis  Bar strength = max of (tangential force of m=2 component)/(radial force)]


Mechanisms:

Gas accretion favor bar because it increases disk mass.

 Bar pattern speed increases.

(Because accretion increases central mass and supply angular momentum)


Ωp increases faster than (inner mass)1/2 ≑ Ω

 Rco decreases with time

(Rco/Rdisk decreases also, because Rdisk itself increases by gas accretion)






Out-of-plane accretion (3D models)

inclined accretion onto a small region of the disk



Results:


- qualitatively the same as 2D

- stronger bar than 2D (because accretion itself introduces non-axisymmetry)


◆ Berentzen et al. (2004,MN,347,220)

Once formed bars generally weaken spontaneously or disappear by gas inflow.

But we see many barred galaxies in nearby universe.

How to regenerate bars?


One Possibility = Tidal interaction
Host galaxy ( = disk + halo, both active) in which the first spontaneous bar has disappeared is perturbed by a companion ( = a rigid Plummer sphere).
1) Host galaxy models are bar-unstable and form bars.

These bars weakened (but not completely vanished)  Isolated model I0

2) The final state of this model was run with gas replaced by stars

Weak steady bar lasts for a long time  Isolated model I1


Interaction = planar, prograde passage. At the closest passage, the weak bar points toward the companion.
Results:

- All the interaction models failed to regenerate a bar in I0 (model including gas disk)

- Sufficiently strong interractions regenerated a long-lasting bar in I1 (purely stellar disk)
Reason: Gas accumulates to the center in I0, which inhibits bar formation.

(Also, angular momentum decrease due to interaction is small in I0)


Implications: Tidally regenerated bars will explain some of high-z and local bars
◆ Immeli et al. (2004,AA,413,547)

Galaxy collapse simulations including dissipative gas and star formation

- Galaxy starts as a (gas+dark matter) sphere( dark matter makes a static potential)

- Gas enters simulation volume at |z|=15 kpc with centrifugally balanced state

and a infall rate of 120 solar mass/yr for initial 1 Gyr

(Gas modeled as an ensemble of inelastically colliding clouds)



Dependence on dissipation efficiency:

- Strong dissipation model  rapid build-up of disk  clump formation in early gas-rich disk  clumps merge and fall to center (‘bulge’ formation)

Star formation rate is high, and attains a peak at bugle formation epoch

- Weak dissipation model  disk settles slowly  bar instability in largely stellar disk in late epoch  bar becomes shorter and pattern speed increases (because of mass accretion to center) (bar always ends at co-rotation)


Paucity of high-redshift barred galaxies (Abraham & Merrifield) may be explained,

Because model bar takes a long time before it appears, and once formed it weakens by the gas accumulation to the center.



Star formation rate is low, and attains the second peak when bar induces gas inflow to the center. (First peak = disk formation)


1   2   3


Verilənlər bazası müəlliflik hüququ ilə müdafiə olunur ©anasahife.org 2016
rəhbərliyinə müraciət