Ultraviolet Fe II emission in z ~ 2 quasars

We present spectra of six luminous quasars at z ~ 2, covering rest wavelengths 1600-3200 A. The fluxes of the UV Fe II emission lines and Mg II 2798 doublet, the line widths of Mg II, and the 3000 A luminosity were obtained from the spectra. These quantities were compared with those of low-redshift quasars at z = 0.06 - 0.55 studied by Tsuzuki et al. In a plot of the Fe II(UV)/Mg II flux ratio as a function of the cental black hole mass, Fe II(UV)/Mg II in our z ~ 2 quasars is systematically greater than in the low-redshift quasars. We confermed that luminosity is not responsible for this excess. It is unclear whether this excess is caused by rich Fe abundance at z ~ 2 over low-redshift or by non-abundance effects such as high gas density, strong radiation field, and high microturbulent velocity.

💡 Research Summary

The paper presents a focused spectroscopic study of six luminous quasars at a redshift of approximately two, aiming to investigate the ultraviolet (UV) Fe II emission and its relationship to the Mg II λ2798 doublet, black‑hole mass, and continuum luminosity. The authors obtained rest‑frame spectra covering 1600–3200 Å, a wavelength interval that encompasses the strongest Fe II UV multiplets as well as the Mg II line. By fitting a Fe II UV template (derived from Vestergaard & Wilkes 2001) and modeling the Mg II profile with a two‑component Gaussian, they measured the integrated fluxes of both species, the full width at half maximum (FWHM) of Mg II, and the monochromatic luminosity at 3000 Å (L₃₀₀₀).

Using the virial method, the Mg II FWHM together with the empirically calibrated BLR radius–luminosity relation (R_BLR ∝ L₃₀₀₀^0.5) yielded estimates of the central black‑hole masses (M_BH) in the range 10⁸–10⁹ M_⊙. The key diagnostic ratio, Fe II(UV)/Mg II, was then compared with a low‑redshift control sample (z = 0.06–0.55) compiled by Tsuzuki et al. (2006), for which identical analysis procedures were applied.

The comparison reveals a systematic offset: the z ≈ 2 quasars exhibit Fe II(UV)/Mg II ratios that are on average 1.5–2 times larger than those of the low‑z objects, even after accounting for differences in continuum luminosity. Statistical tests (Pearson correlation and linear regression) confirm that the ratio does not correlate significantly with L₃₀₀₀, indicating that the excess is not a simple luminosity effect.

Two broad classes of explanations are discussed. The first invokes an enhanced iron abundance at z ≈ 2 relative to magnesium. Since iron is predominantly produced in Type Ia supernovae, a higher Fe/Mg ratio would require either an elevated Ia supernova rate or a rapid enrichment history that allows sufficient time for the delayed iron contribution (≈ 1 Gyr after star formation). At a cosmic age of roughly 3 Gyr, such enrichment is plausible but would demand a specific star‑formation and chemical‑evolution scenario.

The second class attributes the elevated ratio to non‑abundance physical conditions within the broad‑line region (BLR). High gas density (n_H ≫ 10⁹ cm⁻³) enhances collisional excitation of Fe II, while a strong ionizing radiation field (high ionization parameter U) can increase the population of the relevant Fe II levels. Moreover, a large microturbulent velocity (v_turb ≈ 100 km s⁻¹) broadens line profiles, reduces line saturation, and effectively raises the Fe II emissivity without a corresponding increase in Mg II. These effects can produce a higher Fe II(UV)/Mg II ratio even at solar metallicity.

The authors acknowledge several limitations. The sample size (six objects) is modest, limiting statistical robustness. The Fe II template, while widely used, simplifies the complex multiplet structure and may not capture subtle variations in physical conditions. Additionally, only UV diagnostics are employed; complementary measurements of other metallic lines (e.g., C IV, Si IV, Fe II in the near‑infrared) would help disentangle abundance from excitation effects.

Future work is suggested along several lines: expanding the high‑redshift sample to improve statistical significance; obtaining higher‑resolution spectra (R > 5000) to resolve individual Fe II components; conducting multi‑wavelength campaigns that include near‑infrared Fe II bands (e.g., λ ≈ 9200 Å) and optical metal lines; and performing detailed photoionization modeling that simultaneously varies metallicity, density, ionization parameter, and microturbulence. Such efforts would enable a more definitive assessment of whether the observed Fe II excess reflects genuine chemical evolution (i.e., a higher Fe/Mg ratio at earlier cosmic times) or is primarily driven by changes in the physical state of the BLR gas.

In summary, this study provides the first systematic evidence that UV Fe II emission relative to Mg II is enhanced in quasars at z ≈ 2 compared with their low‑redshift counterparts. The result points to either a rapid iron enrichment episode or a shift toward denser, more turbulent BLR environments during the epoch of peak quasar activity. Clarifying the dominant cause will require larger, multi‑band datasets and sophisticated modeling, but the present work establishes a solid observational foundation for probing the co‑evolution of supermassive black holes, their host galaxies, and the chemical enrichment of the early universe.