We present high signal‐to‐noise ratio spectrophotometric observations of seven luminous H ii galaxies. The observations have been made with the use of a double‐arm spectrograph which provides spectra with a wide wavelength coverage, from 3400 to 10 400 Å free of second‐order effects, of exactly the same region as that of a given galaxy. These observations are analysed applying a methodology designed to obtain accurate elemental abundances of oxygen, sulphur, nitrogen, neon, argon and iron in the ionized gas. Four electron temperatures and one electron density are derived from the observed forbidden line ratios using the five‐level atom approximation. For our best objects, errors of 1 per cent in te([O iii]), 3 per cent in te([O ii]) and 5 per cent in te([S iii]) are achieved with a resulting accuracy of 7 per cent in total oxygen abundances, O/H. The ionization structure of the nebulae can be mapped by the theoretical oxygen and sulphur ionic ratios, on the one side, and the corresponding observed emission line ratios, on the other – the η and η′ plots. The combination of both is shown to provide a means to test photoionization model sequences presently applied to derive elemental abundances in H ii galaxies.

We propose a methodology to perform a self‐consistent analysis of the physical properties of the emitting gas of H ii galaxies adequate to the data that can be obtained with the 21st century technology. This methodology requires the production and calibration of empirical relations between the different line temperatures that should supersede currently used ones based on very simple, and poorly tested, photoionization model sequences. As a first step to reach these goals, we have obtained simultaneous blue to far red long‐slit spectra with the William Herschel Telescope (WHT) of three compact H ii galaxies selected from the Sloan Digital Sky Survey (SDSS) Data Release 2 (DR2) spectral catalogue using the INAOE Virtual Observatory superserver. Our spectra cover the range from 3200 to 10 500 Å, including the Balmer jump, the [O ii] λλ 3727, 29 Å lines, the [S iii] λλ 9069, 9532 Å doublet as well as various weak auroral lines such as [O iii] λ 4363 Å and [S iii] λ 6312 Å. For the three objects, we have measured at least four line temperatures, T([O iii]), T([S iii]), T([O ii]) and T([S ii]), and the Balmer continuum temperature T(Bac). These measurements and a careful and realistic treatment of the observational errors yield total oxygen abundances with accuracies between 5 and 9 per cent. These accuracies are expected to improve as better calibrations based on more precise measurements, both on electron temperatures and densities, are produced. We have compared our obtained spectra with those downloaded from the SDSS DR3 finding a satisfactory agreement. The analysis of these spectra yields values of line temperatures and elemental ionic and total abundances which are in general agreement with those derived from the WHT spectra, although for most quantities they can only be taken as estimates since, due to the lack of direct measurements of the required lines, theoretical models had to be used whose uncertainties are impossible to quantify. The ionization structure found for the observed objects from the O+/O2+ and S+/S2+ ratios points to high values of the ionizing radiation, as traced by the values of the ‘softness parameter’η which is less than 1 for the three objects. The use of line temperatures derived from T([O iii]) based on current photoionization models yields for the two highest excitation objects, much higher values of η which would imply lower ionizing temperatures. This is, however, inconsistent with the ionization structure as probed by the measured emission‐line intensities. Finally, we have measured the T(Bac) for the three observed objects and derived temperature fluctuations. Only for one of the objects, the temperature fluctuation is significant and could lead to higher oxygen abundances by about 0.20 dex.

We analyzed the evolution of the metallicity of the gas with the redshift for a sample of AGNs in a very wide redshift range (0 < z < 4) using ultraviolet emission-lines from the narrow-line regions (NLRs) and photoionization models. The new index C43=log[(C iv+C iii])/He ii] is suggested as a metallicity indicator for AGNs. Based on this indicator, we confirmed the no metallicity evolution of NLRs with the redshift pointed out by previous works. We found that metallicity of AGNs shows similar evolution than the one predicted by cosmic semi-analytic models of galaxy formation set within the Cold Dark Matter merging hierarchy (for z 3). Our results predict a mean metallicity for local objects in agreement with the solar value (12+log(O/H)=8.69).This value is about the same that the maximum oxygen abundance value derived for the central parts of local spiral galaxies. Very low metallicity log(Z/Z ⊙ ) ≈ −0.8 for some objects in the range 1.5 < z < 3 is derived.

We examine the relation between oxygen abundances in the narrow-line regions (NLRs) of active galactic nuclei (AGNs) estimated from the optical emission lines through the strong-line method (the theoretical calibration of Storchi-Bergmann et al. 1998), via the direct T e -method, and the central intersect abundances in the host galaxies determined from the radial abundance gradients. We found that the T e -method underestimates the oxygen abundances by up to ∼2 dex (with average value of ∼ 0.8 dex) compared to the abundances derived through the strong-line method. This confirms the existence of the so-called "temperature problem" in AGNs. We also found that the abundances in the centres of galaxies obtained from their spectra trough the strong-line method are close to or slightly lower than the central intersect abundances estimated from the radial abundance gradient both in AGNs and Star-forming galaxies. The oxygen abundance of the NLR is usually lower than the maximum attainable abundance in galaxies (∼2 times the solar value). This suggests that there is no extraordinary chemical enrichment of the NLRs of AGNs.