Views on the Abundance of Nitrogen with GALEX

1
Views on the Abundance of Nitrogen with GALEX
Ryan Mallery

• University of California, Los Angeles

R. Michael Rich (UCLA), L. Kewley (IFA), S. Salim
(NOAO), C.Tremonti (Steward Observatory), GALEX
science team
2
Outline

• Brief introduction to nitrogen abundances
• GALEX galaxy sample, SFR, M estimates
• Nitrogen abundance diagnostics
• Results

3
Nitrogen ProductionPrimary or Secondary

• Primary Nitrogen
• nitrogen that is produced independent of a stars
  initial metallicity
• Secondary Nitrogen
• nitrogen that is metallicity dependent
• Both nitrogen yields are dependent on stellar
  mass and metallicity.
• Both primary and secondary nitrogen produced by
  intermediate and high mass stars

Chiappini et al. (2006)
4
Nitrogen Abundances in Galaxies

• At O/H
• At 7.6of 0.3 dex

• What Causes the Scatter ?
• Variation of the IMFs
• Galactic Winds
• Mixing timescale/abundance anisotropy
• Time delay between the release of N and O,
  varying star formation history models

Van Zee Haynes 2006
5
Mixing Timescales

• Izotov et al. 2005,
• Te abundances of metal poor SDSS galaxies.
• claims that the scatter in N/O is due to
  chemical inhomogeneities
• Winds from WR rich in primary N, while
  concentrated, the nitrogen rich wind can increase
  N/O by 0.15 dex
• Once wind mixes N/O will only globally increase
  by 0.03 dex

12 log O/H
6
Mixing Timescales
Chiappini et al. 2005 chemical evolution models
of differing star formation histories can
reproduce the nitrogen and oxygen abundances
without primary N from massive stars.
12 log O/H
7
Time Delay Scenario

• In a starburst, the massive stars evolve first
  and release oxygen into the ISM massive stars,
  raising O/H and lowering N/O.
• Nitrogen is released much later in intermediate
  mass stars.
• N/O depends on star formation history of a galaxy.

8
Time Delay Scenario

• The time delay scenario is too simplistic.
• Galaxies have more complex star formation
  histories than bursts followed by long quiescent
  periods.
• Galaxies with current SFR less than their past
  average SFR will have enhanced nitrogen van Zee
  et al. 2006

9

• Liang et al. 2006
• SDSS DR4 N/O strong line abundances
• For a given oxygen abundance, galaxies with
  higher N/O ratios have lower EWH?
• N/O -- depends of star formation history of a
  galaxy

10
Galaxy Sample

• 8,745 galaxies in GALEX SDSS DR4 spectroscopic
  sample
• 645 sq. deg. overlap
• SDSS Spectroscopy
• resolution ????? 2000
• wavelength range 380-920 nm
• GALEX
• FUV (1350 - 1800 Å)
• NUV (1800 - 2800 Å)
• Selection Criteria
• NUV or FUV detection
• 5???????????????????
• ?????????????????????????????
• r-band spectral flux 20 of total r-band flux

11
GALEX NGSGALEX SDSS
12
Determining SFR and M

• GALEXSDSS broadband SED
• FUV,NUV, ugriz
• -- fit SED to 105 Bruzual Charlot models
  parameterized by
• attenuation Charlot Fall (2000)
• star formation history
• Age of galaxy
• -- Generate Bayesian estimates
• ????of each fit -- weight assigned to the
  parameters that model.
• M, SFR, ?V, etc.
• see Salim et al 2005, 2007

Bruzual Charlot, 2003
13
SFR GALEX vs H?

• UV GALEX SDSS
• Average SFR over the last 100Myr.
• H? Brinchmann et al 2005
• Star formation over last
• 10 Myr

14
Nitrogen Abundance Diagnostics

• Te or recombination line diagnostics for low
  metallicity galaxies.
• Only one strong line diagnostic for galaxies
  where weak recombination lines or auroral lines
  are not detected.

15
Nitrogen Abundance Strong Line Diagnostic

• Log N/O Log NII6584/OII3727 0.307 -
  0.02 Log TNII - 0.726/TNII
• from Pagel et al (1992)
• Empirical calibration from Te abundances
• Applicability to high metallicities is unknown.

• TNII from Cloudy
• photoionization models
• TNII 500K uncertainty

• TNII .6065.1600 log R23 .1878 (log R23 )2
  .2803 (log R23)3
• from Thurston, Edmonds Henry (1996)

16
Strong line diagnostic vs Te diagnostic

• Only 33 galaxies have
• 3? OIII4363 detections
• Strong line diagnostic overestimates the Te
  diagnostic by 0.1 dex with 0.05 dex of scatter.

• Mean Te error
• 0.06 dex for sample of 33 galaxies
• Mean strong line error
• 0.05 dex for the sample of 33 galaxies
• .17 dex for entire sample

17
N/O Star formation History
18
Oxygen Diagnostics

• Degeneracies
• O3N2 Pettini Pagel et al 2005
• OIII/?? / NII/??
• Not corrected for ionization
• M91 McGaugh et al (1991)
• R23 -- O/H
• R23 -- TNII --N/0
• T04 Tremonti et al 2004
• O/H
• Completely primary
• O/H 8.2 N/H (O/H)2
• Completely secondary

19
N/O vs O/H
red SFR/M -9.1 blue SFR/M -9.1

• The N/O ratio of a galaxy is on average lower
  for galaxies with higher SFR/M
• The large errors on N, coupled with the
  degeneracies hinder a definitive conclusions.
• IMF variations, and mixing timescales are not
  ruled out.

20
Mass Color
21
Te Diagnostic

• N/O vs O/H
• Shows large scatter
• Small sample all with high EW??
• Only slight indication that N/O decreases for
  increase of EW?? or SRR/M

22
UV Luminous Galaxies

• LFUV 1010
• LFUV/ru2 109
• GALEX has detected 232 in SDSS
• 53 match emission line criteria.

UVLGs are oxygen deficient, and normal to
enhanced nitrogen -- are galactic
winds removing oxygen?
23
Summary

• At a given metallicity, a galaxy will on average
  have a higher N/O ratio the lower its specific
  star formation rate.
• Oxygen diagnostics that depend on NII
• Should take this into account if they want to be
  accurate.
• The N/O strong line diagnostic lacks precision,
  average error of 0.17 dex, and its accuracy is
  unknown.
• Te abundances show a large scatter in N/O.
• caused by WR winds?
• More accurate and precise diagnostics to
  determine N/O are needed.

24
ENDING
25
(No Transcript)
26
(No Transcript)
27
CNO cycle