Astron. Astrophys. 354, 17-27 (2000)

1. Introduction

Quasars are among the most luminous objects known in the universe. Due to their high luminosity they can be observed easily to redshifts of z5 or even more. Currently, nearly 100 quasars with redshifts larger than z = 4 are known and recently the first quasar with z = 5.0 was discovered by Fan et al. (1999) in late 1998. The redshift range of z3 corresponds to an epoch when the universe had an age of approximately 10 % of its current age (assuming H_o=65 km s^-1 Mpc^-1, =0.3, =0.7; cf. Carroll et al. 1992; Carlberg et al. 1999; Perlmutter et al. 1999). Therefore, quasars provide a powerful probe to study the early history of the universe. Especially, with respect to dating the first star formation epoch quasars have gained increasing interest. Their prominent emission-line spectrum can be used a diagnostic tool to estimate the enrichment of the gas with heavy elements due to violent star formation (for a review see Hamann & Ferland 1999).

Recent studies of quasars at high redshift (z3) provide evidence for significantly enhanced metallicities up to an order of magnitude compared to solar metallicities (cf. Hamann & Ferland 1992, 1993; Ferland et al. 1996; Korista et al. 1996; Pettini 1999). These derived high metallicities require a violent star formation phase.

The prominent spectra of quasars provide much information on the physical conditions and the elemental abundances of the gas. But the emission line strength is quite insensitive to metallicity effects. For example, the Ly/ CIV1549 ratio is nearly independent to heavy element abundances (Hamann & Ferland 1999). However, there are several emission line ratios which are sensitive to relative abundance ratios and these can be used to put constraints on both the metallicity and evolution parameters.

The key of the approach using those line ratios to estimate the metallicity are the different time scales of the enrichment of gas with -elements like oxygen, neon, or magnesium, carbon as an additional primary element, nitrogen as a secondary element, or iron. While -elements are produced predominantly in massive stars on short time scales the production of C as a primary and N as a secondary element is ascribed to intermediate mass stars which evolve on significant longer time scales. The dominant source of iron is given by SN Ia explosions which are ascribed stars of intermediate mass (cf. Wheeler et al. 1989).

Previous investigations to estimate the abundances in BELR gas based on the measurement of several generally weak intercombination lines like NIV]1486, OIII]1663, NIII]1750, and CIII]1909 (cf. Shields 1976; Davidson 1977; Baldwin & Netzer 1978; Osmer 1980; Gaskell et al. 1981; Uomoto 1984). Furthermore, and even more serious these lines are subject to collisional de-excitation for densities larger than n_e = 10¹⁰cm^{- 3}. In spite of these limitations and of the current assumption of higher densities of the BLR gas (Rees et al. 1989; Ferland et al. 1992; Peterson 1993; Baldwin et al. 1995) the results of the early estimates indicate larger than solar metallicities for the BELR gas. Since the investigations of Hamann & Ferland (1992, 1993), Osmer et al. (1994), and Ferland et al. (1996) alternative line ratios gain increasing interest. The NV1240 emission is of special interest since this emission line of a secondary element is generally stronger than expected in the spectra of high redshift quasars in the frame work of standard photoionization models. Hamann & Ferland (1992, 1993) showed that NV1240/CIV1549 and NV1240/HeII1640 can be used as robust metallicity indicators. However, NV1240 and HeII1640 are generally difficult to measure since both lines are severely blended by Ly and CIV1549, OIII]1663, respectively.

In Sect. 2 we described the observations and the data analysis of our sample of 16 quasars with redshift z3. In Sect. 3 we present the results of the analysis of the emission line spectra. We derived the elemental abundance of the line emitting gas based on the line ratios of several diagnostical emission lines. While the line ratio Ly/CIV1549 is nearly independent of metallicity effects NV1240/CIV1549 and NV1240/HeII1640 provide suitable information to estimate the elemental abundance of the line emitting gas (e.g. Ferland et al. 1996; Korista et al. 1997; Hamann & Ferland 1999). Using these emission-line ratios we derived elemental abundances of Z8 for the BELR gas of the quasar sample we observed. The results are discussed and are compared with recent studies (e.g. Ferland et al. 1996; Hamann & Ferland 1999). Our results provide further evidence that the first violent star formation epoch might start at a redshift of z_f 6 to 10.

© European Southern Observatory (ESO) 2000

Online publication: January 31, 2000
helpdesk.link@springer.de