We analyse absorption lines in optical and ultraviolet spectra of the brigh
t (V = 15.8, z = 2.19) QSO HS 1103+6416. High-resolution (FWHM = 8 km s(-1)
) optical spectra have been obtained with the Keck 10 m telescope in the ra
nge from 3180 to 5780 Angstrom. Ultraviolet observations in the range from
1150 to 3280 Angstrom were performed with the FOS and the GHRS onboard the
Hubble Space Telescope (HST). In this paper we concentrate our discussion o
n a complex Lyman limit system (LLS) at z = 1.89, Absorption lines by carbo
n, silicon and aluminum in the optical spectra reveal a complex velocity st
ructure with at least 11 components spanning a velocity range of 200 km s(-
1). From the Lyman limit in the ultraviolet spectra we derive a total neutr
al hydrogen column density of log N(HI) = 17.46 cm(-2).
Column densities of heavy elements in the individual components were derive
d by Voigt profile fitting. The eleven components can be subdivided roughly
into three groups: Components 2, 3 and 6 with radial velocities v = -129..
. -95 km s(-1) with low ionization (L), components 4, 5, 7, 8 (v = -75... 2) with intermediate ionization (I), and components 1, 9, 10, 11 (v = -129,
+34.., +57) with high ionization (H). In order to study the ionization and
abundances in these systems we compare the observed column densities with
photoionization models. The observed absorption in the optical data can be
explained by individual clouds with slightly varying metal abundances photo
ionized by slightly different radiation fields. Highly ionized components f
avour the extragalactic radiation field as calculated by Haardt & Madau (19
96) while the components of low and intermediate ionization an better repro
duced with a harder ionizing radiation field. Obviously local sources like
stars can therefore be excluded as the main ionizing sources.
Abundances in components L and I appear to be slightly different from those
in the high ionization component H.
In L and I we find roughly [C/H] = -0.9 while H has [C/H] = -1.2, consisten
t with the expectation that in a galaxy or groups of galaxies the abundance
s in the higher ionized 'Halo' component are lower. The relative element ab
undances are also different. While in components L and I [Si/C] approximate
to 0.2, barely significant, and [S/C] and [O/C] approximate to 0 within th
e uncertainties, component H shows [Si/C] = 0.5 and in addition [O/C] and [
S/C] = 0.4 (both from HST spectra). [AVC] measurable only in L and I is alw
ays approximate to 0.
The tendency of enhanced a element (O, Si, S) abundances at low C abundance
is consistent with what is known from nucleosynthesis theory (SNII dominan
t at the beginning of galactic evolution), from metal deficient stars in ou
r galaxy and from QSO absorption line systems, If all components were ioniz
ed by the same radiation field the relative overabundances of O and S in th
e highly ionized components would be even larger.
We show that HS 1103+6416 will offer in the future fur the first time the p
ossibility to measure the cosmic He abundance at high redshift. Detailed ca
lculations of He I absorption using the multicomponent model which explains
the metal lines shows consistency with the observed first seven series mem
bers of the He I 584, 537, 522 Angstrom... series for a helium abundance Y
= 0.24, the expected cosmic He abundance from Big Bang nucleosynthesis modi
fied by stellar nucleosynthesis at similar to 1/10 solar metallicity.
The presence of OI and possibly O VI absorption cannot be explained by our
photoionization models and might hint at the existence of additional mainly
neutral components with relatively low H I column density and further ioni
zation mechanisms like, e.g., collisional ionization.