1. Metallicity: A Smoking Gun for Gas Flows in Mergers
Lisa Kewley
U. Hawaii
Margaret Geller (SAO),
Betsy Barton (UC Irvine)

2. Summary Motivation Metallicity diagnostics
Luminosity-Metallicity Relation
Central Star Formation
Blue Bulges
Merger Models
Conclusions & Future Work

3. Motivation Effect of mergers on metallicity? unknown
Predicted gas flows?
elusive Fe Si Mg
Chandra X-ray spectra : Metallicity map of the hot ISM of the Antennae, where red, green, and blue indicate emission by Fe, Si, and Mg, respectively. The continuum was subtracted by estimating the contribution of the overall best-fit two-component thermal bremsstrahlung model in the two bands at 1.4–1.65 and 2.05–3.50 keV, where no strong lines are observed. Pointlike sources were excluded, following Fabbiano et al. (2003b). The Fe-L image was adaptively smoothed to a significance between 3 and 5 σ, and the same scales were applied to the other line images.
Fabbiano et al. (2004)

4. Galaxy Pairs Field Galaxies 502 galaxies from the CfA redshift catalog
(v > 2300 km/s, Dv < 1035 km/s, DD < 77 h-1 Mpc)
(Barton et al. 2000)
Nuclear spectra for ~200 galaxies in pairs
Field Galaxies
198 galaxies from the CfA redshift catalog
full range in Hubble type & Magnitudes in CfA survey
(Jansen et al. 2000)
In normal spiral galaxies, nuclear spectra can be corrected (albeit with some uncertainty) for aperture effects, if we know what HUbble type and luminosity a galaxy has, so that we can correct for metallicity gradients. However with galaxy pairs, we may run into a
further complication: dynamical effects of tidally-driven gas flows. In the following slides we investigate this effect.

5. Metallicity Diagnostics
“R23”
Kewley & Dopita (2002, ApJS, 142, 35)
Also: Pagel (1979), McCall et al. (1985), ...,
Skillman et al. (1989), McGaugh (1991),...,
Zaritsky et al. (1994), Charlot (2001), ...

6. Galaxy Pairs Luminosity-metallicity Relation 1. shifts for close pairs
Although a few studies of individual galaxies using integral field or HI velocity maps show evidence for gas flows in merging galaxies, the effect on the metallicity gradient, and hence on the nuclear metallicities of galaxy pairs is
To investigate the effect of possible dynamical
Kewley, Geller, & Barton
(2005, AJ, 131, 2004)

7. Luminosity Effect? Need 1-2 Mag rise Kewley, Geller, & Barton
wwide pairs
close pairs
Need 1-2 Mag rise
Smoothed particle hydrodynamic simulations such as the one shown here predict that the tidal forces induce gravitational torques which remove angular momentum from the gas. The gravitational potential of the galaxy also becomes distorted which then affects
the motions of the gas. The gas rapidly accumulates into the spiral arms of the galaxy, forming shocked, dense filaments. In a dissipative medium, these filiments allow rapid streaming of the gas down the spiral arms and into the nuclear region. Simulations predict that gas infall occurs at an average rate of 7 solar masses per year after the initial close pass, and that the infall continues until 100 Myr. If we calculate the amount of infall and timescale required to dilute the nuclear metallicities by ~0.2 dex, we require an infall
timescale of 9-90 Myr, consistent with the merger simulations.
NFGS
cz < 2300 km/s
Kewley, Geller, & Barton
(2006, AJ, 131, 2004)

8. Luminosity Effect? No shift in upper bound : MB Negative shift in
wwide pairs
No shift in upper
bound : MB
Negative shift in
right bound: metallicity
close pairs
Smoothed particle hydrodynamic simulations such as the one shown here predict that the tidal forces induce gravitational torques which remove angular momentum from the gas. The gravitational potential of the galaxy also becomes distorted which then affects
the motions of the gas. The gas rapidly accumulates into the spiral arms of the galaxy, forming shocked, dense filaments. In a dissipative medium, these filiments allow rapid streaming of the gas down the spiral arms and into the nuclear region. Simulations predict that gas infall occurs at an average rate of 7 solar masses per year after the initial close pass, and that the infall continues until 100 Myr. If we calculate the amount of infall and timescale required to dilute the nuclear metallicities by ~0.2 dex, we require an infall
timescale of 9-90 Myr, consistent with the merger simulations.
NFGS
cz < 2300 km/s
Kewley, Geller, & Barton
(2006, AJ, 131, 2004)

9. Luminosity Effect? R-Band Luminosity-metallicity Relation
1. still shifts for close pairs
Although a few studies of individual galaxies using integral field or HI velocity maps show evidence for gas flows in merging galaxies, the effect on the metallicity gradient, and hence on the nuclear metallicities of galaxy pairs is
To investigate the effect of possible dynamical
Kewley, Geller, & Barton
(2006, AJ, 131, 2004)

10. Galaxy Pairs metallicity effect? gas infall? Luminosity-metallicity
Relation
1. shifts for close pairs
metallicity effect?
gas infall?
Although a few studies of individual galaxies using integral field or HI velocity maps show evidence for gas flows in merging galaxies, the effect on the metallicity gradient, and hence on the nuclear metallicities of galaxy pairs is
To investigate the effect of possible dynamical
Kewley, Geller, & Barton
(2006, AJ, 131, 2004)

11. Central Burst Strength, SR(t)
Barton, Geller & Kenyon (2003):
Stellar population synthesis models + colors + EWs
assuming 2 populations (old & young)
SR(t) = current fraction of R-band light
from young burst

12. Central Burst Strength
Barton, Geller &
Kenyon (2003)

13. Galaxy Pairs Luminosity-metallicity Relation 1. shifts for close pairs
2. correlated with central burst
strength
Kewley, Geller, & Barton
(2006, AJ, 131, 2004)

14. Galaxy Pairs Luminosity-metallicity Relation 1. shifts for close pairs
2. correlated with central burst
strength
Kewley, Geller, & Barton
(2006, AJ, 131, 2004)

15. Blue Bulges (B-R) = (B-R)outer - (B-R)inner where
Kannappan et al. (2003): “blue bulge parameter”
(B-R) = (B-R)outer - (B-R)inner
where
(B-R)outer = (B-R) at 75% light radius
(B-R)inner = (B-R) at 1/2 light radius

16. Galaxy Pairs Luminosity-metallicity Relation 1. shifts for close pairs
Blue Bulges
Luminosity-metallicity
Relation
1. shifts for close pairs
2. correlated with central burst
strength
3. correlated with blue bulges
Kewley, Geller, & Barton
(2006, AJ, 131, 2004)

17. Metallicity Gradient Keck LRIS Spectroscopy Kewley, Geller, & Barton
(2005, AJ, submitted)

18. Metallicity Gradient Keck LRIS Spectroscopy See also: Li Hsin Chien’s
Poster
Kewley, Geller, & Barton
(2005, AJ, submitted)

19. Merger Scenario Images Courtesy Chris Mihos
Smoothed particle hydrodynamic simulations such as the one shown here predict that the tidal forces induce gravitational torques which remove angular momentum from the gas. The gravitational potential of the galaxy also becomes distorted which then affects
the motions of the gas. The gas rapidly accumulates into the spiral arms of the galaxy, forming shocked, dense filaments. In a dissipative medium, these filiments allow rapid streaming of the gas down the spiral arms and into the nuclear region. Simulations predict that gas infall occurs at an average rate of 7 solar masses per year after the initial close pass, and that the infall continues until 100 Myr. If we calculate the amount of infall and timescale required to dilute the nuclear metallicities by ~0.2 dex, we require an infall
timescale of 9-90 Myr, consistent with the merger simulations.
Images Courtesy Chris Mihos

20. Galaxy Pairs Luminosity-metallicity Relation 1. shifts for close pairs
2. correlated with central burst
strength
3. correlated with blue bulges
Evidence for Gas Infall
Kewley, Geller, & Barton
(2006, AJ, 131, 2004)

21. Metallicity Gradients
Smoothed particle hydrodynamic simulations such as the one shown here predict that the tidal forces induce gravitational torques which remove angular momentum from the gas. The gravitational potential of the galaxy also becomes distorted which then affects
the motions of the gas. The gas rapidly accumulates into the spiral arms of the galaxy, forming shocked, dense filaments. In a dissipative medium, these filiments allow rapid streaming of the gas down the spiral arms and into the nuclear region. Simulations predict that gas infall occurs at an average rate of 7 solar masses per year after the initial close pass, and that the infall continues until 100 Myr. If we calculate the amount of infall and timescale required to dilute the nuclear metallicities by ~0.2 dex, we require an infall
timescale of 9-90 Myr, consistent with the merger simulations.

22. Metallicity Gradients
Smoothed particle hydrodynamic simulations such as the one shown here predict that the tidal forces induce gravitational torques which remove angular momentum from the gas. The gravitational potential of the galaxy also becomes distorted which then affects
the motions of the gas. The gas rapidly accumulates into the spiral arms of the galaxy, forming shocked, dense filaments. In a dissipative medium, these filiments allow rapid streaming of the gas down the spiral arms and into the nuclear region. Simulations predict that gas infall occurs at an average rate of 7 solar masses per year after the initial close pass, and that the infall continues until 100 Myr. If we calculate the amount of infall and timescale required to dilute the nuclear metallicities by ~0.2 dex, we require an infall
timescale of 9-90 Myr, consistent with the merger simulations.
Kennicutt, Bresolin & Garnett (2003)

23. Merger Scenario . Iono et al. (2004): Simulations predict:
1. Gas inflow rate ~ 7 Mo/yr
2. Gas flows within 1st 100 Myr
but before disk merger .
Smoothed particle hydrodynamic simulations such as the one shown here predict that the tidal forces induce gravitational torques which remove angular momentum from the gas. The gravitational potential of the galaxy also becomes distorted which then affects
the motions of the gas. The gas rapidly accumulates into the spiral arms of the galaxy, forming shocked, dense filaments. In a dissipative medium, these filiments allow rapid streaming of the gas down the spiral arms and into the nuclear region. Simulations predict that gas infall occurs at an average rate of 7 solar masses per year after the initial close pass, and that the infall continues until 100 Myr. If we calculate the amount of infall and timescale required to dilute the nuclear metallicities by ~0.2 dex, we require an infall
timescale of 9-90 Myr, consistent with the merger simulations.

24. Merger Scenario . Iono et al. (2004): Simulations predict:
1. Gas inflow rate ~ 7 Mo/yr
2. Gas flow within 100 Myr
but before disk merger .
Smoothed particle hydrodynamic simulations such as the one shown here predict that the tidal forces induce gravitational torques which remove angular momentum from the gas. The gravitational potential of the galaxy also becomes distorted which then affects
the motions of the gas. The gas rapidly accumulates into the spiral arms of the galaxy, forming shocked, dense filaments. In a dissipative medium, these filiments allow rapid streaming of the gas down the spiral arms and into the nuclear region. Simulations predict that gas infall occurs at an average rate of 7 solar masses per year after the initial close pass, and that the infall continues until 100 Myr. If we calculate the amount of infall and timescale required to dilute the nuclear metallicities by ~0.2 dex, we require an infall
timescale of 9-90 Myr, consistent with the merger simulations.

25. Merger Scenario . Assuming: How much infalling gas is required?
1. Gas inflow rate ~ 7 Mo/yr
2. Normal Spiral Metallicity gradient .
How much infalling gas is required?
Smoothed particle hydrodynamic simulations such as the one shown here predict that the tidal forces induce gravitational torques which remove angular momentum from the gas. The gravitational potential of the galaxy also becomes distorted which then affects
the motions of the gas. The gas rapidly accumulates into the spiral arms of the galaxy, forming shocked, dense filaments. In a dissipative medium, these filiments allow rapid streaming of the gas down the spiral arms and into the nuclear region. Simulations predict that gas infall occurs at an average rate of 7 solar masses per year after the initial close pass, and that the infall continues until 100 Myr. If we calculate the amount of infall and timescale required to dilute the nuclear metallicities by ~0.2 dex, we require an infall
timescale of 9-90 Myr, consistent with the merger simulations.

26. Merger Scenario . Assuming: How much central dilution is required?
1. Gas inflow rate ~ 7 Mo/yr
2. Normal Spiral Metallicity gradient .
How much central dilution is required?
50-60%
Merger models predict: 60% infall
Smoothed particle hydrodynamic simulations such as the one shown here predict that the tidal forces induce gravitational torques which remove angular momentum from the gas. The gravitational potential of the galaxy also becomes distorted which then affects
the motions of the gas. The gas rapidly accumulates into the spiral arms of the galaxy, forming shocked, dense filaments. In a dissipative medium, these filiments allow rapid streaming of the gas down the spiral arms and into the nuclear region. Simulations predict that gas infall occurs at an average rate of 7 solar masses per year after the initial close pass, and that the infall continues until 100 Myr. If we calculate the amount of infall and timescale required to dilute the nuclear metallicities by ~0.2 dex, we require an infall
timescale of 9-90 Myr, consistent with the merger simulations.

27. Merger Scenario . Assuming: . How long will it take to infall?
1. Gas inflow rate ~ 7 Mo/yr
2. Normal Spiral Metallicity gradient
3. Central Gas Mass Mo
ave v. high . .
How long will it take to infall?
Smoothed particle hydrodynamic simulations such as the one shown here predict that the tidal forces induce gravitational torques which remove angular momentum from the gas. The gravitational potential of the galaxy also becomes distorted which then affects
the motions of the gas. The gas rapidly accumulates into the spiral arms of the galaxy, forming shocked, dense filaments. In a dissipative medium, these filiments allow rapid streaming of the gas down the spiral arms and into the nuclear region. Simulations predict that gas infall occurs at an average rate of 7 solar masses per year after the initial close pass, and that the infall continues until 100 Myr. If we calculate the amount of infall and timescale required to dilute the nuclear metallicities by ~0.2 dex, we require an infall
timescale of 9-90 Myr, consistent with the merger simulations.

28. Merger Scenario . Assuming: . How long will it take to infall?
1. Gas inflow rate ~ 7 Mo/yr
2. Normal Spiral Metallicity gradient
3. Central Gas Mass Mo . .
How long will it take to infall?
9x x107 years
Merger models predict: within 1x108 years
Smoothed particle hydrodynamic simulations such as the one shown here predict that the tidal forces induce gravitational torques which remove angular momentum from the gas. The gravitational potential of the galaxy also becomes distorted which then affects
the motions of the gas. The gas rapidly accumulates into the spiral arms of the galaxy, forming shocked, dense filaments. In a dissipative medium, these filiments allow rapid streaming of the gas down the spiral arms and into the nuclear region. Simulations predict that gas infall occurs at an average rate of 7 solar masses per year after the initial close pass, and that the infall continues until 100 Myr. If we calculate the amount of infall and timescale required to dilute the nuclear metallicities by ~0.2 dex, we require an infall
timescale of 9-90 Myr, consistent with the merger simulations.

29. New Theoretical Models
Josefa Perez et al. (2006, astro-ph/ )
Chemical evolution model (Scannappieco et al. 2005)
L-CDM model (GADGET-2; Springel & Hernquist 2003)
Lower mean central (O/H) in pairs from inflows +
Smoothed particle hydrodynamic simulations such as the one shown here predict that the tidal forces induce gravitational torques which remove angular momentum from the gas. The gravitational potential of the galaxy also becomes distorted which then affects
the motions of the gas. The gas rapidly accumulates into the spiral arms of the galaxy, forming shocked, dense filaments. In a dissipative medium, these filiments allow rapid streaming of the gas down the spiral arms and into the nuclear region. Simulations predict that gas infall occurs at an average rate of 7 solar masses per year after the initial close pass, and that the infall continues until 100 Myr. If we calculate the amount of infall and timescale required to dilute the nuclear metallicities by ~0.2 dex, we require an infall
timescale of 9-90 Myr, consistent with the merger simulations.

30. Conclusions Close galaxy pairs have lower than field
central metallicities
“Smoking gun” for gas infall during merger?
Central metallicity correlates with:
central burst strength
blue bulges
Timescale consistent with current
merger simulations

31. Future Directions Keck LRIS spectra of matched pair members
Merger simulations of metallicity gradients

32. Starburst99-Mappings On-Line L. Kewley & C. Leitherer
Available Now!
Starburst99-Mappings On-Line L. Kewley & C. Leitherer
Starburst99-Mappings Interface:
Mappings Interface:
Pre-run model grids
Interactive web form to run models

33. Motivation Effect of mergers on metallicity is unknownSi Mg
Chandra X-ray spectra : Metallicity map of the hot ISM of the Antennae, where red, green, and blue indicate emission by Fe, Si, and Mg, respectively. The continuum was subtracted by estimating the contribution of the overall best-fit two-component thermal bremsstrahlung model in the two bands at 1.4–1.65 and 2.05–3.50 keV, where no strong lines are observed. Pointlike sources were excluded, following Fabbiano et al. (2003b). The Fe-L image was adaptively smoothed to a significance between 3 and 5 σ, and the same scales were applied to the other line images.
Fabbiano et al. (2004)

34. Metallicity Diagnostic Comparisons
Kewley & Ellison
(2005)

35. Metallicity: Strong lines vs Auroral Lines
Garnett, Kennicutt, Bresolin

36. Classification Scheme

37. Galaxy Pairs Luminosity-metallicity Relation 1. shifts for close pairs1’
Luminosity-metallicity
Relation
1. shifts for close pairs
2. correlated with central burst
strength
3. correlated with blue bulges
Evidence for Gas Infall
Kewley, Geller, & Barton
(2005, AJ, submitted)

38. GOODS Survey: 0.3 < z< 1
Kobulnicky & Kewley
(2004, ApJ, 617,240)

39. Metallicity - [NII]/Ha
Pettini & Pagel (2004)

40. Metallicity Diagnostics
1. Theoretical - photoionization models
e.g., McGaugh (1991), Kewley & Dopita (2002),
Tremonti et al. (2004)
2. Empirical - fit to Te metallicities
e.g., Pilyugin (2000), Pettini & Pagel (2004)
Combination - fit to Te method +
theoretical metallicities
e.g., Denicolo, Terlevich & Terlevich (2002)

41. Metallicity - [OIII]/Hb,[NII]/Ha
Pettini & Pagel (2004)

42. Metallicity Diagnostic Comparisons
Kewley & Ellison
(2005, in prep)

43. Metallicity: Strong lines vs Auroral Lines
Garnett, Kennicutt, Bresolin

44. GOODS Survey: 0.3 < z< 1
Kobulnicky & Kewley (2004, ApJ, 617, 240)

45. Auroral Line Saturation
Stasinska (2002,2005)