Metallicities of star forming galaxies atz

As reviewed by Pettini (2006), the problem of determining the metallicities in star-forming galaxies at z = 2- 3 has been tackled with a variety of methods, using

8

_ 1

1 1 1 1 1 1

1 1

Ha

-

6

-

[N II] X6548 [N II] X6583 [S ii]

U6716, 6731 "

2

i j; i i

_

0

/ LA rv

6400

6500 6600 6700

6800

Wavelength (A)

8

_ 1

1 i 1 i 1 i

1 1

-

Ha

6

-

r 11

-

4

-

[N II] X6548 [N II] X6583 [S I

U6716, 6731 -

2

-

I I 1

0

___

-----— N - ^ T^r^ay^j^^^1^ lj j. ■■■ ■.. .--1---

6400 6500 6600 6700 6800

Wavelength (A)

Figure 20.4. Composite spectra of BX galaxies at z — 2.2 from the survey by Erb et al. (2006a). Upper panel: galaxies brighter than Ks = 20 have a mean ratio [N ii]/ Ha = 0.25, which indicates an oxygen abundance of 12 + log (O/H) = 8.56, or about four fifths Solar, if the local calibration of the N2 index with (O/H) applies to these galaxies. Lower panel: galaxies fainter than Ks = 20 have [N ii]/Ha = 0.13 and 12 + log (O/H) = 8.39, or about half Solar.

6400 6500 6600 6700 6800

Wavelength (A)

Figure 20.4. Composite spectra of BX galaxies at z — 2.2 from the survey by Erb et al. (2006a). Upper panel: galaxies brighter than Ks = 20 have a mean ratio [N ii]/ Ha = 0.25, which indicates an oxygen abundance of 12 + log (O/H) = 8.56, or about four fifths Solar, if the local calibration of the N2 index with (O/H) applies to these galaxies. Lower panel: galaxies fainter than Ks = 20 have [N ii]/Ha = 0.13 and 12 + log (O/H) = 8.39, or about half Solar.

nebular emission, interstellar absorption and stellar features in the UV, including both wind and photospheric lines. These different approaches have given roughly concordant answers (to within a factor of ~2) in the few cases so far when they could be cross-checked, but a realistic assessment of the current situation would be to say that we are still some way off from developing reliable metallicity indicators that are easily applicable to high-z galaxies. The diagnostic which has had the widest application up to now is the N2 index - the ratio of the intensities of the [N ii] A6583 and Ha emission lines - which was most recently calibrated by Pettini & Pagel (2004) in terms of the oxygen abundance in nearby H ii regions. One of the advantages of this index is that it relies on the ratio of two emission lines that are closely spaced in wavelength, thereby disposing of the need for accurate flux and reddening determinations (see Figure 20.4).

When Erb et al. (2006a) applied the local N2 calibration to their sample of 87 z ~ 2 galaxies grouped in six bins of stellar mass, a clear mass-metallicity relation emerged (see Figure 20.5). There are several points that can be made from consideration of Figure 20.5. First, these results certainly dispel any naive ideas that high-redshift galaxies may be generally metal-poor - on the contrary, 85%

Figure 20.5. The stellar mass-metallicity relation for star-forming galaxies at z ~ 2 (large circular points with error bars) and in the nearby (z ~ 0.1) Universe from the SDSS survey (numerous small grey points). The large circular points correspond to the sample of 87 UV-selected galaxies at z — 2 grouped by Erb et al. (2006a) into six bins according to their stellar mass. The triangles show the mean metallicities of the SDSS galaxies in the same mass bins as the z — 2 galaxies. The vertical error bar in the bottom right-hand corner indicates the uncertainty in the calibration of the N2 index used to deduce the oxygen abundances of both sets of galaxies. Since the N2 index saturates near Solar metallicity (indicated by the horizontal dotted line), the offset between the present-day and high-redshift relations is best derived by consideration of the lower-metallicity bins. At a given assembled stellar mass, galaxies at z = 2 appear to be about a factor of two lower in metallicity than today. Figure reproduced from Erb et al. (2006a).

Figure 20.5. The stellar mass-metallicity relation for star-forming galaxies at z ~ 2 (large circular points with error bars) and in the nearby (z ~ 0.1) Universe from the SDSS survey (numerous small grey points). The large circular points correspond to the sample of 87 UV-selected galaxies at z — 2 grouped by Erb et al. (2006a) into six bins according to their stellar mass. The triangles show the mean metallicities of the SDSS galaxies in the same mass bins as the z — 2 galaxies. The vertical error bar in the bottom right-hand corner indicates the uncertainty in the calibration of the N2 index used to deduce the oxygen abundances of both sets of galaxies. Since the N2 index saturates near Solar metallicity (indicated by the horizontal dotted line), the offset between the present-day and high-redshift relations is best derived by consideration of the lower-metallicity bins. At a given assembled stellar mass, galaxies at z = 2 appear to be about a factor of two lower in metallicity than today. Figure reproduced from Erb et al. (2006a).

of the BX galaxies in the Erb et al. sample have (O/H) > |(O/H)©. Second, it is hard to assess how metal-rich the most massive galaxies actually are, for the reason discussed by Fabio Bresolin at this meeting: the inherent uncertainties in the nebular abundance scale at supersolar metallicities. Specifically, the N2 index saturates near Solar oxygen abundance, as is also shown by the cloud of SDSS points in Figure 20.5. Possibly, in this regime the C iv and Si iv P-Cygni lines from luminous OB stars, which are commonly seen in the UV spectra of these galaxies, would be a more reliable tell-tale sign of supersolar abundances (Rix et al. 2004), but no clear-cut such cases have been identified yet at high or low z (see Chapter 6 in these proceedings). In any case, the Solar or slightly subsolar metallicities of

"v

Redshift

Figure 20.6. Abundances of Zn in 87 DLAs, from the compilation by Kulkarni et al. (2005) (open symbols), which brings together the results of several surveys for DLAs in optically selected QSO samples, and from the recent survey of 20 CORALS (radio-selected) QSOs by Akerman et al. (2005) (filled symbols). Triangles denote upper limits in DLAs whose Zn ii XX2026, 2062 lines remain undetected.

"v

Redshift

Figure 20.6. Abundances of Zn in 87 DLAs, from the compilation by Kulkarni et al. (2005) (open symbols), which brings together the results of several surveys for DLAs in optically selected QSO samples, and from the recent survey of 20 CORALS (radio-selected) QSOs by Akerman et al. (2005) (filled symbols). Triangles denote upper limits in DLAs whose Zn ii XX2026, 2062 lines remain undetected.

even the most massive BX galaxies in Figure 20.5 are not necessarily inconsistent with the proposal that they are the progenitors of today's massive ellipticals, since the latter exhibit supersolar metallicities only in their inner regions, whereas the metallicity measures at high z refer to the integrated light of a whole galaxy. It will be of great interest to investigate the presence of abundance gradients in BX galaxies with spatially resolved observations, once integral field spectrographs with adaptive-optics correction are fully operative (Forster-Schreiber et al. 2006; Law et al. 2006).

A final point concerning the mass-metallicity relation is that, even within the uncertainties of the metallicity calibrations, there appears to be a real offset between the relation today and that at z = 2, in the sense that galaxies of a given stellar mass were a factor of ~ 2 less metal-enriched in the distant past (10 Gyr ago, or three quarters of the age of the Universe in today's 'consensus' cosmology) than they are now. This is weak evolution indeed, as argued earlier, and consistent with the shallow metallicity-redshift gradients deduced by observations of galaxies at intermediate redshifts (e.g. Savaglio et al. 2005, Maier et al. 2006).

Are DLAs relevant to a meeting on the metal-rich Universe? The answer would at first seem to be negative: it has been a source of some concern over the last decade that these absorption systems, which in principle should give us an unbiased measure of the progress of cosmic chemical evolution (e.g. Pei & Fall 1995), are generally metal-poor at all redshifts sampled (see Figure 20.6). Indeed, one could argue that DLAs offer us one of the best means to identify the most metal-poor gas at high redshifts and, with highly precise measures of its chemical composition, provide

Was this article helpful?

0 0

Post a comment