Mon. Not. R. Astron. Soc. 386, 22092220 (2008) doi:10.1111/j.1365-2966.2008.13186.

Metal abundances at z < 1.5: new measurements in sub-damped Lyman

absorbers

C. Peroux,1 J. D. Meiring,2 V. P. Kulkarni,2 P. Khare,3 J. T. Lauroesch,4 G. Vladilo5

and D. G. York6
1 Observatoire Astronomique de Marseille Provence, Marseille, Laboratoire dAstrophysique de Marseille, UMR6110, CNRS/Universite de Provence, France
2 Department of Physics and Astronomy, University of South Carolina, Columbia, USA
3 Department of Physics, Utkal University, Bhubaneswar, India
4 Department of Physics and Astronomy, University of Louisville, USA
5 Osservatorio di Trieste, Trieste, Italy
6 Department of Astronomy and Astrophysics and Enrico Fermi Institute, University of Chicago, Chicago, USA

Accepted 2008 March 4. Received 2008 March 3; in original form 2008 January 11

Downloaded from http://mnras.oxfordjournals.org/ by guest on September 8, 2016

ABSTRACT
Damped Lyman systems (DLAs) and sub-DLAs seen towards background quasars provide the
most detailed probes of elemental abundances. Somewhat paradoxically these measurements
are more difficult at lower redshifts due to the atmospheric cut-off, and so a few years ago
our group began a program to study abundances at z < 1.5 in quasar absorbers. In this
paper, we present new Ultraviolet and Visual Echelle Spectrograph (UVES) observations of
six additional quasar absorption-line systems at z < 1.5, five of which are sub-DLAs. We find
solar or above solar metallicity, as measured by the abundance of zinc, assumed not to be
affected by dust, in two sub-DLAs: one, towards Q01380005 with [Zn/H] = +0.28 0.16;
the other towards Q2335+1501 with [Zn/H] = +0.07 0.34. Relatively high metallicity was
observed in another system: Q01230058 with [Zn/H] = 0.45 0.20. Only for the one
DLA in our sample, in Q04491645, do we find a low metallicity, [Zn/H] = 0.96 0.08.
We also note that in some of these systems large relative abundance variations from component
to component are observed in Si, Mn, Cr and Zn.
Key words: galaxies: abundances intergalactic medium quasars: absorption lines
quasars: general.

of the age of the Universe. The gas phase abundances, measured

1 INTRODUCTION
spectroscopically, are often depleted relative to the true abundances
Damped Lyman systems (hereafter DLAs) seen in absorption in because some of the atoms are incorporated into dust grains. For
the spectra of background quasars, select galaxies over all red- this reason, the abundances of metal-rich/dusty absorbers can only
shifts independent of their intrinsic luminosity. They have hydrogen be derived from volatile elements, which show little affinity with the
column densities, logN(H I)  20.3. DLAs are also the major con- dust, such as sulfur and zinc. The only volatile, nearly undepleted
tributors to the neutral gas density, H I , in the Universe at high element accessible at low redshifts with ground-based instruments
redshifts (Peroux et al. 2003a; Prochaska, Herbert-Fort & Wolfe and with usually unsaturated lines that often lie outside the Lyman
2005) but slightly lower column density systems are also believed forest is zinc (Pettini et al. 1997). Indeed, lines of other elements
to contribute to the H I mass (Peroux et al. 2005). These latter sys- that are normally not highly depleted in the interstellar medium, in
tems are referred to in the community as sub-Damped Lyman particular oxygen, nitrogen and sulfur, but they are often saturated
systems (sub-DLAs) and were defined by Peroux et al. (2003a) to and are not accessible at the low redshifts we study here, except
have 19.0  logN(H I)  20.3. Furthermore, both DLAs and sub- from space.
DLAs offer a direct probe of element abundances over >90 per cent Most cosmic chemical evolution models (e.g. Malaney &
Chaboyer 1996; Pei, Fall & Hauser 1999) predict the global inter-
 Based on observations collected during programme ESO 78.A-0646 at stellar metallicity to rise from almost 1/1000 solar at high redshift to
the European Southern Observatory with UVES on the 8.2-m KUEYEN nearly solar at z = 0. The present-day mass-weighted mean interstel-
telescope operated at the Paranal Observatory, Chile. lar metallicity of local galaxies is, in fact, nearly solar (e.g. Kulka-
E-mail: [email protected] rni & Fall 2002). DLAs and sub-DLAs are expected to follow this


C 2008 The Authors. Journal compilation 
C 2008 RAS
2210 C. Peroux et al.
behaviour if they trace an unbiased sample of galaxies selected only known to have DLAs or sub-DLAs at 0.6 < z < 1.5 with N(H I)
by N(H I). There has been considerable debate about whether or measurement from Hubble Space Telescope (HST).
not high N(H I) quasar absorbers actually show this predicted rise The data were reduced using the most recent version of the UVES
up to solar metallicity. This ambiguity is mainly due to the small pipeline in MIDAS (uves/2.1.0 flmidas/1.1.0). Master bias and flat
number of measurements available for study at z < 1.5, which cor- images were constructed using calibration frames taken closest in
responds to 70 per cent of the age of the Universe. Measurements time to the science frames. The science frames were extracted with
of the metallicities of galaxies detected in emission in this so-called the optimal option. The blue part of the spectra was checked order
redshift desert are extremely sparse, although this redshift interval by order to verify that all were properly extracted. The resulting
covers a significant look-back time. Therefore, information gained spectra were combined, weighting each spectrum by signal-to-noise
in absorption is of utmost importance at these epochs. The problem ratio, corrected to vacuum heliocentric reference and normalized.
at z < 1.5 is that (i) the Zn II lines lie in the ultraviolet (UV) for z < To perform the abundance studies, the spectra were divided into 50
0.6 and in the blue at 0.6 < z < 1.5 and (ii) the H I Lyman lines 75 regions, and each region was normalized using cubic spline
lie in the UV for z < 1.5. Even at these low redshifts, DLAs appear functions of the orders of 1 to 5 as the local continuum. The final
to have low metallicities (e.g. Kulkarni et al. 2005; Meiring et al. spectra have a resolution R = / 45 000 and are rebinned
2006; Peroux et al. 2006b). to 0.05 /pixel to obtain a significantly improved signal-to-noise
Furthermore, a missing metals problem is known to exist at high ratio. These rebinned spectra were used for all measurements: Voigt
redshifts (e.g. Bouche, Lehnert & Peroux 2005, 2006; Pettini 2006). profile fits, Apparent Optical Depth (AOD) measures and equivalent
Indeed, no more than 3040 per cent of the metals expected to width measurements.
be produced by z = 2.5 are actually observed if one sums up the One of the objects in our programme, Q0217+0144, yielded a

Downloaded from http://mnras.oxfordjournals.org/ by guest on September 8, 2016

metals in the DLAs (which are less common and only 1/10 solar), spectrum with too low signal-to-noise ratio to be useful and there-
the Lyman forest lines (which are much more common but only fore is not included in our sample. This object appears to be vari-
1/1000 solar) and the Lyman-break galaxies (which are relatively able, Bergeron & DOdorico (1986) measured 13.8 AB magnitude at
rare). Where are the missing metals? Some have suggested that 6000 in 1984 August, Blades et al. (1985) B > 18.5, Pettini
warm/hot gas in galactic haloes represents a plausible contribution et al. (1983) report B = 14.516.5 and Bolton & Wall (1969) B >
to the missing metals problem (Ferrara & Scannapieco 2005; Fox 19.5. While at the time of submission of the observing proposal our
et al. 2007; Sommer-Larsen & Fynbo 2008). Another possibility is best estimate for the magnitude was g 16, it dimmed by at least
that the more metal-rich DLAs are dustier, obscure the background three magnitudes by the time the observations took place. This
quasars more, and are therefore under-represented in the present strong variation in magnitude suggests that the object might be a
samples. The results of the search for DLAs towards radio-selected BL Lac.
quasars suggest that dust obscuration might be modest (Ellison et al.
2001a), but the sample is still small to derive firm conclusions.
3 A N A LY S I S
Dust effects can be important if the extinction increases with metal
column density as found in interstellar clouds (Vladilo & Peroux
3.1 Methodology
2005; Vladilo et al. 2006). Indeed, the dust content increases with
metallicity (Vladilo 2004) and dust is expected to play a significant In the following, each section title includes the name of the quasar,
role at z < 1.5 if the metallicity increases with cosmic time. A way the value of the H I column density, the absorber redshift and the
to find metal-rich absorbers minimizing dust effects is to search for number of components. The error estimates of the N(H I) mea-
quasar absorbers with relatively low H I column density. In fact, surements are provided in Table 1. They are derived from Space
recent data are suggesting that metal-rich systems can be found Telescope Imaging Spectrograph (STIS) observations by Rao,
more easily in these absorbers than in classical DLAs (e.g. Pettini Turnshek & Nestor (2006) except for Q2335+1501 where Ad-
et al. 2000; Jenkins et al. 2005). In particular, the growing sample vanced Camera for Surveys (ACS) prism spectra (prog ID 10556;
of observations of sub-DLAs includes some high-metal abundance PI: D. Turnshek; S. Rao et al., private communication) were used.
(e.g. Khare et al. 2004; Peroux et al. 2006a; Meiring et al. 2007, As shown in Meiring et al. (2008), the single component fits used
2008). are degenerate in logN(H I) in relatively low resolution, low signal-
In this paper, we present high-resolution observations of six to-noise ratio HST data, leading to artificially high H I column den-
quasars taken with Ultraviolet and Visual Echelle Spectrograph sities. This makes our measured abundances, [X/H], conservative
(UVES) at Very Large Telescope (VLT). These observations in- in the sense that they are slightly lower than the actual ones. The
crease the sample of low-redshift sub-DLAs observed at high- absorber redshifts reported in Table 1 are from Rao et al. (2006).
resolution to determine metallicity using dust-free estimates based The metal column densities were estimated as in Meiring et al.
on zinc. The paper is structured as follows. Target selection strategy (2006, 2007, 2008) and Peroux et al. (2006a,b) by fitting mutliple-
and details of observations are provided in Section 2. In the third component Voigt profiles to the observed absorption lines using the
section, we detail the analysis and the profile fit of each of our quasar program FITS6P (Welty, Hobbs & York 1991) which evolved from
absorbers, including the total abundances with respect to solar. Fi- the code used by Vidal-Madjar et al. (1977). FITS6P minimizes the
nally, the last section analyses these results in terms of metallicity 2 between the data and the theoretical Voigt profiles convolved
and dust content. with the instrumental profile. The same fitting procedure was used
for all the systems. The number of components were kept to a min-
imum but future higher resolution observations might require more
2 DATA R E D U C T I O N
components than fitted in this work as indicated by relatively large
The targets for our observing programme were generally selected b values. Basically, the velocity and Doppler b parameters of the
from the Sloan Digital Sky Survey (SDSS) sample of quasars, aug- various components were estimated from the Mg I, Mg II and Fe II
mented with few objects from the literature, with Mg II absorbers lines when available and then kept fixed for the remaining lines,


C 2008 The Authors. Journal compilation 
C 2008 RAS, MNRAS 386, 22092220
 Metal abundances at z < 1.5 2211
Table 1. Journal of observations. N(H I) column densities are in cm2 .

Quasar SDSS ID g maga z emb z absc log N(H I) Observation date c (nm) Texp (s)

Q01230058 SDSS J012303.22005818.9 18.7 1.549 1.4094 20.08+0.10

0.08 2006 October 01 520 22700
Q0132+0116 SDSS J013233.90+011607.2 18.8 1.786 1.2712 19.70+0.08
0.10 2006 October 03 520 3 4800 + 11166
Q01380005 SDSS J013825.53000534.5 18.8 1.341 0.7821 19.81+0.06
0.11 2006 October 12/13/19 346+564 35200
Q0153+0009 SDSS J015318.19+000911.3 17.9 0.838 0.7714 19.70+0.08
0.10 2006 October 04 346+564 25400
Q04491645d 18.0 2.679 1.0072 20.98+0.06
0.07 04/14/20 Oct 2006 390+564 35200
Q2335+1501 SDSS J233544.19+150118.3 18.2 0.790 0.6798 19.70+0.30
0.30 2006 October 03/11/13 346+564 44480
a Magnitudes are SDSS g band when available and V band otherwise.
bz is from the SDSS when available.
em
cz
abs is from Rao et al. (2006).
d Q04491645 was not observed by SDSS.

1.5 1.5

1.0 1.0

0.5 0.5

0.0
AlIII1854 0.0
CrII2056

Downloaded from http://mnras.oxfordjournals.org/ by guest on September 8, 2016

1.5 1.5

1.0 1.0

0.5 0.5

0.0
AlIII1862 0.0
FeII2249
1.5 1.5

1.0 1.0

0.5 0.5
Normalised Flux

0.0
MnII2576 0.0
FeII2260
1.5 1.5

1.0 1.0

0.5 0.5

0.0
SiII1808 0.0
FeII2344
1.5 1.5

1.0 1.0

0.5 0.5

0.0
TiII1910 0.0
FeII2374
1.5 1.5

1.0 1.0

0.5 0.5

0.0
ZnII2026 0.0
FeII2382
1.5
-250 -125 0 125 250
1.0 -1
Radial Velocity (km s )
0.5

0.0
ZnII2062
-250 -125 0 125 250
-1
Radial Velocity (km s )

Figure 1. Voigt profile fit of the z abs = 1.4094 sub-DLA towards Q01230058 (see Table 2). The dashed vertical lines in the Zn II 2026 and 2062 panels
indicate the position of the Mg I 2026 and Cr II 2062 components, respectively. The spectra are rebinned to 0.05 /pixel.

allowing for variations from metal species to metal species in N tions to the total column densities are negligible and therefore not
only. This procedure is standard practice, and is appropriate be- included in the total sum. Lower limits in a given saturated com-
cause the b values are not expected to show systematic trends with ponent of the profile fit were estimated from the fit. In case the
atomic mass because they are dominated by bulk motions rather components with the strongest lines of other species are noisy, we
than thermal broadening. The resulting error estimates for the ve- derive upper limits to the total column densities. Upper limits to the
locities are 0.51.0 km s1 and less than 1 km s1 for the b values. total column densities deduced from non-detection were derived at
All the transitions observed were used together to get the column 3 using the local signal-to-noise ratio estimate.
density for each species. In particular, the very weak Fe II 2249,
2260 lines, even when not detected, constrain the values of N(Fe II).
In all cases, the Mg I 2026.5 contribution to the Zn II 2026.1 line 3.2 Individual systems
was estimated using the component parameters for Mg I derived
from the Mg I 2852 profile and using fosc (Mg I 2852)/f osc (Mg I 3.2.1 Q01230058; logN(H I) = 20.08; z abs = 1.4094;
2026) = 32. In Figs 1-6, for Zn II 2026.14, the nearby Mg I 10 components
2026.48 line is noted by vertical dashed lines. Likewise for Cr II A number of features are detected over some or all of the profile
2062.23 in the plots of Zn II 2062.66. In all cases, both pairs components: Al III 1854, 1862; Si II 1808; Mn II 2576; Cr II
of ions are fitted simultaneously, thus including the contribution of 2056; Fe II ( 2249, 2260, 2344, 2374, 2382) and Zn II 2026,
the weaker line that can be estimated given the high resolution of 2062. Ti II 1910 is covered but not detected. Based on the SDSS
the data. The atomic data were adopted from Morton (2003). All DR3 spectra, we note that the Al III 1854 is blended with the
wavelengths reported here are truncated vacuum values. Mn II 2594 line from a lower redshift strong Mg II absorber (z abs =
In Tables 27, we give the resulting column densities. The few 0.723). In this system, it is interesting to note that the column density
weak components, in which the column densities could not be well ratios of Si, Mn, Cr and Zn vary from one component to another.
constrained due to higher noise are marked with . Their contribu- By comparison, the ratio of column densities of Al III and Fe II is


C 2008 The Authors. Journal compilation 
C 2008 RAS, MNRAS 386, 22092220
2212 C. Peroux et al.
1.5 1.25

1.0 1.00

0.5 0.75

AlIII1854 CrII2056
0.0 0.50
1.5 1.25

1.0 1.00

0.5 0.75

AlIII1862 FeII2260
0.0 0.50
1.25 1.5

1.00 1.0
Normalised Flux
0.75 0.5
MnII2576 FeII2344
0.50 0.0
1.25 1.5

1.00 1.0

0.75 0.5
TiII1910 FeII2374
0.50 0.0
1.25 1.5

1.00 1.0

0.75 0.5
ZnII2026 FeII2382
0.50 0.0
1.25 1.5

1.00 1.0

0.75 0.5
ZnII2062

Downloaded from http://mnras.oxfordjournals.org/ by guest on September 8, 2016

0.50 0.0
FeII2586
-250 -125 0 125 250 -250 -125 0 125 250
-1 -1
Radial Velocity (km s ) Radial Velocity (km s )

Figure 2. Voigt profile fit of the z abs = 1.2712 sub-DLA towards Q0132+0116 (see Table 3). Fe II 2344 is blended with the Mg II 2796 line from a known
Mg II system z abs = 0.9050 (shaded in grey). The peaks at v = +100 km s1 in Ti II 1910 and Mn II 2576 and v = +190 km s1 in Mn II 2576 are due to
bad sky subtraction in the reduction process.

1.5 1.25

1.0 1.00

0.5 0.75

AlIII1854 CrII2056
0.0 0.50
1.5 1.5

1.0 1.0

0.5 0.5

0.0
AlIII1862 0.0
FeII2600
1.5 1.5

1.0 1.0
Normalised Flux

0.5 0.5

0.0
MgII2796 0.0
MgI2852
1.5 1.25

1.0 1.00

0.5 0.75

MgII2803 MnII2606
0.0 0.50
1.25 1.25

1.00 1.00

0.75 0.75
ZnII2026 TiII3383
0.50 0.50
1.25
-250 -125 0 125 250

1.00 Radial Velocity (km s -1)

0.75
ZnII2062
0.50
-250 -125 0 125 250
-1
Radial Velocity (km s )

Figure 3. Voigt profile fit of the z abs = 0.7821 sub-DLA towards Q01380005 (see Table 4).

about 1.2 in all components, close to the mean found by Meiring Zn II 2026, 2062 are covered but not detected. Upper limits from
et al. (2007, 2008). Assuming no variations in the radiation field, non-detections imply log N(Mn II) <11.42, log N(Cr II) <12.13 and
relative abundances of the noted elements seem to be variable from log N(Zn II) <11.76. The Mg I 2852 line fall in a spectral gap and
component to component. The results of the profile fitting analysis the Mg I 2026 line does not provide restrictive limits. The results
are summarized in Table 2, while Fig. 1 illustrates the fits. of the profile fitting analysis are summarized in Table 3, while Fig. 2
illustrates the fits.

3.2.2 Q0132+0116; logN(H I) = 19.70; z abs = 1.2712;

11 components 3.2.3 Q01380005; logN(H I) = 19.81; z abs = 0.7821;
five components
In this system, we clearly detect the following species: Al III
1854, 1862 and Fe II ( 2260, 2344, 2374, 2382, 2586). However, The following species are detected and fitted in this sub-DLA:
Mn II ( 2576, 2594, 2606); Ti II 1910; Cr II 2056, 2062 and Mg I 2852; Mg II 2796, 2803 and Zn II 2026, 2062. Only


C 2008 The Authors. Journal compilation 
C 2008 RAS, MNRAS 386, 22092220
 Metal abundances at z < 1.5 2213
1.5 1.5

1.0 1.0

0.5 0.5

0.0
AlIII1854 0.0
CrII2056
1.5 1.5

1.0 1.0

0.5 0.5

0.0
AlIII1862 0.0
MgI2852
1.5 1.5

1.0 1.0

Normalised Flux 0.5 0.5

0.0
TiII3242 0.0
MgII2796
1.5 1.5

1.0 1.0

0.5 0.5

0.0
TiII3383 0.0
MgII2803
1.5 1.5

1.0 1.0

0.5 0.5

0.0
ZnII2026 0.0
SiII1808
1.5
-300 -150 0 150 300

1.0 Radial Velocity (km s -1)

0.5

Downloaded from http://mnras.oxfordjournals.org/ by guest on September 8, 2016

0.0
ZnII2062
-300 -150 0 150 300
-1
Radial Velocity (km s )

Figure 4. Voigt profile fit of the z abs = 0.7714 sub-DLA towards Q0153+0009 (see Table 5). Obvious apparent emission features are sky subtraction
artefacts.

1.5 1.5

1.0 1.0

0.5 0.5

0.0 AlIII1854 0.0 MgI2852

1.5 1.5

1.0 1.0

0.5 0.5

0.0 AlIII1862 0.0 TiII1910

1.5 1.5

1.0 1.0

0.5 0.5

0.0 FeII2344 0.0 MnII2576

1.5 1.5
Normalised Flux

1.0 1.0

0.5 0.5

0.0 FeII2374 0.0 MnII2594

1.5 1.5

1.0 1.0

0.5 0.5

0.0 FeII2382 0.0 MnII2606

1.5 1.5

1.0 1.0

0.5 0.5

0.0 FeII2586 0.0 CrII2056

1.5 1.5

1.0 1.0

0.5 0.5

0.0 CoII1941 0.0 ZnII2026

1.5 1.5

1.0 1.0

0.5 0.5

0.0 SiII1808 0.0 ZnII2062

-250 -125 0 125 250 -250 -125 0 125 250

Radial Velocity (km s -1) Radial Velocity (km s -1)

Figure 5. Voigt profile fit of the z abs = 1.0072 DLA towards Q04491645 (see Table 6). The areas marked in grey are forest blends.

the saturated Fe II 2600 line is detected (the other Fe II lines

3.2.4 Q0153+0009; logN(H I) = 19.70; z abs = 0.7714;
fall in a spectral gap), the profile fittings of these lines provide
18 components
lower limits to the column densities. The following lines: Al III
1854, 1862; Si II 1808; Cr II 2056; Mn II 2606 (the oth- The following species are clearly detected: Mg I 2852 and Mg II
ers lines for Mn II fall in a spectral gap); Ti II ( 3073, 3230, 2796, 2803. In addition, Al III 1854, 1862; Cr II 2056; and
3242, 3384) and Na I ( 3303, 3304) are covered but not de- Ti II 3242, 3383 are all detected in the strongest component at
tected. Upper limits from non-detections are log N(Al III) < 12.28, v = 156.4 km s1 . In this system, Mg I 2852 is also detected and is
log N(Si II) <14.89, log N(Cr II) <12.61, log N(Mn II) <11.84, log particularly saturated at +160 km s1 . The Ti II lines are extremely
N(Ti II) <11.48 and log N(Na I) < 13.05. The results of the profile strong, being at least 10 times stronger than in DLAs (Dessauges-
fitting analysis are summarized in Table 4, while Fig. 3 illustrates the Zavadsky, Prochaska & DOdorico 2002; Ledoux, Bergeron &
fits. Petitjean 2002), even though N(H I) is 10 times weaker. Zn II


C 2008 The Authors. Journal compilation 
C 2008 RAS, MNRAS 386, 22092220
2214 C. Peroux et al.
1.25 1.25

1.00 1.00

0.75 0.75
CaII3934 CrII2056
0.50 0.50
1.25 1.5

1.00 1.0

0.75 0.5
FeII2249
MgI2852
Normalised Flux
0.50 0.0
1.25 1.5

1.00 1.0

0.75 0.5
FeII2260
MgII2803
0.50 0.0
1.25 1.25

1.00 1.00

0.75 0.75
ZnII2026 TiII3383
0.50 0.50
1.25
-250 -125 0 125 250

1.00 Radial Velocity (km s -1)

0.75
ZnII2062

Downloaded from http://mnras.oxfordjournals.org/ by guest on September 8, 2016

0.50
-250 -125 0 125 250

Radial Velocity (km s -1)

Figure 6. Voigt profile fit of the z abs = 0.6798 sub-DLA towards Q2335+1501 (see Table 7).

Table 2. Parameters for the Voigt profile fit of the z abs = 1.4094 sub-DLA towards Q01230058 (see Fig. 1). Components marked are not detected in the
corresponding transition. Velocities and b values are in km s1 and column densities are in cm2 .

Vel b N(Al III) N(Si II) N(Mn II) N(Cr II) N(Fe II) N(Zn II)

120.7 9.9 (1.49 0.34)E12 (1.55 1.23)E14 (5.84 0.63)E12

75.0 8.7 (3.73 0.48)E12 (2.41 1.29)E14 (8.25 0.53)E13
58.9 8.6 (5.40 0.57)E12 (2.40 1.28)E14 (8.51 0.55)E13
30.7 9.8 (1.89 0.17)E13 (6.20 1.57)E14 (2.50 0.87)E12 (1.66 0.09)E14
13.7 6.2 (9.15 1.28)E12 (7.04 1.70)E14 (1.56 0.72)E12 (2.76 0.45)E14 (3.82 1.84)E11
0.5 5.7 (5.56 0.98)E12 (2.63 1.48)E14 (5.62 0.95)E13 (6.32 2.03)E11
12.8 11.5 (9.85 0.88)E12 (9.04 1.95)E14 (8.18 1.83)E12 (2.20 0.12)E14 (6.90 2.49)E11
38.2 8.3 (1.78 0.37)E12 (7.47 1.64)E14 (4.96 0.25)E13
56.4 9.9 (1.86 0.37)E12 (1.81 0.10)E13
85.0 6.2 (2.80 0.50)E12

Table 3. Parameters for the Voigt profile fit of the z abs = 1.2712 sub-DLA Table 4. Parameters for the Voigt profile fit of the z abs = 0.7821 sub-DLA
towards Q0132+0116 (see Fig. 2). towards Q01380005 (see Fig. 3).

Velocity b N(Al III) N(Fe II) Vel b N(Mg I) N(Mg II) N(Fe II) N(Zn II)

128.1 4.9 (4.20 0.71)E12 45.3 7.0 (7.29 0.50)E11 >2.20E14 >7.67E13 (1.89 0.22)E12
103.2 10.7 (1.21 0.31)E12 (2.07 0.13)E13 24.1 6.1 (4.68 0.40)E11 >8.59E13 >4.26E14 (1.00 0.19)E12
59.9 10.1 (2.27 0.36)E12 (1.71 0.13)E13 3.3 6.7 (1.35 0.10)E12 >1.47E15 >6.95E14 (1.14 0.40)E12
34.5 10.5 (8.54 0.67)E12 (7.22 0.39)E13 19.7 9.9 (1.54 0.09)E12 >1.32E14 >2.85E14
16.1 9.8 (2.10 0.40)E12 (1.21 0.15)E13 49.4 8.6 (4.72 0.41)E11 >1.79E13 >1.03E13 (8.57 2.03)E11
5.6 10.3 (2.65 0.38)E12 (3.20 0.18)E13
34.2 13.2 (6.13 0.51)E12 (5.36 0.26)E13
60.0 6.8 (3.13 0.44)E12 (4.09 0.34)E13
77.0 10.5 (8.96 0.69)E12 (9.18 0.54)E13 spectra. The results of the profile fitting analysis are summarized in
101.8 3.5 (1.26 0.29)E12 (2.28 0.24)E13 Table 5, while Fig. 4 illustrates the fits.
116.9 8.4 (1.74 0.32)E12 (5.06 0.29)E13

3.2.5 Q04491645; logN(H I) = 20.98; z abs = 1.0072;

13 components
2026, 2062; Si II 1808 and Na I ( 3303, 3304) are covered
but not detected. Upper limits from non-detections lead to log The following features are detected and fitted: Mg I 2852; Al III
N(Zn II) <11.96, log N(Si II) <14.49 and log N(Na I) <13.01. Un- 1854 (partly blended with Lyman forest features) and 1862
fortunately, all the Fe II lines for that system fall in spectral gaps. (clean); Si II 1808; Cr II 2056; Mn II ( 2576, 2594, 2606);
However, several lines of Fe II are detected in the original SDSS DR5 Fe II ( 2344, 2374, 2382, 2586); Zn II 2026, 2062 and Co II


C 2008 The Authors. Journal compilation 
C 2008 RAS, MNRAS 386, 22092220
 Metal abundances at z < 1.5 2215
Table 5. Parameters for the Voigt profile fit of the z abs = 0.7714 sub-DLA towards Q0153+0009 (see Fig. 4). Several lines of Fe II are detected in the original
SDSS spectra. The corresponding EW estimates are reported in Table 8.

Vel b N(Mg I) N(Mg II) N(Al III) N(Cr II) N(Ti II)

245.7 5.2 >1.74E12

234.2 6.0 >1.11E12
180.8 4.9 >1.45E12
169.8 2.8 (7.69 1.28)E10 >3.43E12
131.4 12.8 (6.55 1.96)E10 >5.52E12
108.2 8.9 (8.04 0.27)E11 >4.01E13 -
86.1 8.7 >4.20E12
68.8 9.9 (1.17 0.18)E11 >8.19E12
44.1 5.9 (1.15 0.04)E12 >3.20E13
20.7 9.1 (3.60 0.20)E11 >5.19E13
1.7 7.2 (2.30 0.18)E11 >3.06E13
33.5 9.3 (1.02 0.17)E11 >7.73E12
83.1 8.6 >4.56E12
103.1 9.3 (3.37 1.65)E10 >3.74E12
138.9 8.6 (3.12 0.23)E11 >1.08E13
156.4 8.3 (4.13 0.16)E12 >5.54E15 (4.00 0.50)E12 (1.04 0.13)E12 (6.50 1.52)E12

Downloaded from http://mnras.oxfordjournals.org/ by guest on September 8, 2016

188.8 9.4 (1.59 0.18)E11 >7.30E12
206.6 4.3 (4.90 1.30)E10 >3.39E12

Table 6. Parameters for the Voigt profile fit of the z abs = 1.0072 DLA towards Q04491645 (see Fig. 5). Co II is detected in the first two components (103.6
and 83.7 km s1 ) of the profile with (8.28 3.20)E12 and (1.90 0.33)E13, respectively. Confirmation with another Co II line is needed to show that this is
not an interloper.

Vel b N(Mg I) N(Al III) N(Si II) N(Cr II) N(Mn II) N(Fe II) N(Zn II)

103.6 9.4 (6.26 0.20)E11 (1.28 0.07)E13 (3.00 0.14)E15 (1.48 0.08)E13 (4.45 0.36)E12 (4.86 0.26)E14 (2.99 0.08)E12
83.7 6.8 (2.16 0.14)E11 (5.44 0.50)E12 (4.12 1.85)E14 (4.22 0.58)E12 (1.10 0.23)E12 (1.78 0.15)E14 (6.04 0.53)E11
67.4 6.8 (2.85 0.15)E11 (9.43 0.68)E12 (5.52 2.01)E14 (2.80 0.55)E12 (1.00 0.23)E12 (1.49 0.10)E14
49.3 8.2 (5.82 0.20)E11 (6.37 0.51)E12 (1.92 0.39)E15 (3.80 0.59)E12 (1.11 0.24)E12 (1.69 0.08)E14 (5.55 0.55)E11
33.3 5.2 (1.47 0.13)E11 (4.60 1.91)E14 (4.80 0.25)E13
5.8 13.6 (1.33 0.50)E12 (5.86 0.79)E12
19.6 11.5 (8.21 1.93)E10 (1.79 0.10)E13
36.3 7.7 (3.57 0.52)E12
65.0 7.7 (2.34 1.75)E14 (7.67 0.62)E12
75.1 4.2 (1.68 0.13)E11 (1.67 0.51)E12 (3.22 0.19)E13
88.4 4.9 (2.01 0.13)E11 (1.43 0.32)E12 (4.73 1.87)E14 (1.93 0.49)E12 (4.84 1.88)E11 (8.20 0.45)E13
101.1 5.3 (6.31 1.14)E10 (1.57 0.32)E12 (2.38 1.53)E14 (3.82 0.18)E13
122.1 9.3 (3.89 0.48)E12

Table 7. Parameters for the Voigt profile fit of the z abs = 0.6798 sub-DLA towards Q2335+1501 (see Fig. 6).

Velocity b N(Mg I) N(Mg II) N(Ca II) N(Cr II) N(Fe II N(Zn II)

65.3 7.5 (3.07 0.21)E11 >2.32E14 (1.36 0.63)E11

49.1 6.8 (4.29 0.23)E11 >3.77E14 (3.26 0.65)E11
24.8 7.8 (4.75 0.24)E11 >2.42E13 (3.49 0.69)E11 (4.02 1.33)E12 (1.40 0.27)E14
5.6 8.2 (1.19 0.05)E12 >1.07E14 (5.53 0.79)E11 (1.08 0.28)E14 (8.40 1.44)E11
9.0 7.2 (1.41 0.06)E12 >1.68E13 (8.84 0.86)E11 (3.81 1.41)E12 (2.39 0.29)E14 (9.43 1.39)E11
35.7 7.5 (9.80 0.04)E11 >2.04E15 (3.43 0.46)E11 (1.85 0.27)E14 (5.57 1.27)E11

1941. Lines of Ti II ( 1910, 3073, 3230, 3242, 3384) and Na I while Fig. 5 illustrates the fits. The areas marked in grey are blended
( 3303, 3304) are covered but not detected. Upper limits from with Lyman forest features.
non-detections lead to log N(Ti II) < 11.47 and log N(Na I) < 12.66.
The Mg II doublet for that systems falls in a spectral gap. Interest-
3.2.6 Q2335+1501; logN(H I) = 19.70; z abs = 0.6798;
ingly, the Co II is mutually exclusive of the Si II. The Si II mimics
six components
other elements, but for instance, the strength ratio [Si II/Mn II] is
quite different between 104 and 49 km s1 . Confirmation with We report the detection of the following features: Mg I 2852;
another Co II line is needed to show that this is not an interloper. Mg II 2803; Cr II 2056; Fe II 2249, 2260; Zn II 2026,
The results of the profile fitting analysis are summarized in Table 6, 2062 and Ca II 3934. The Mg II 2796 is falling right at the peak


C 2008 The Authors. Journal compilation 
C 2008 RAS, MNRAS 386, 22092220
2216 C. Peroux et al.
Table 8. Rest-frame equivalent widths and associated errors for the systems in this sample. Values and 3 errors are in units of m. The EW of Ti II 3242
in the sub-DLAs towards Q0153+0009 is 27 7 m. Numbers in brackets are from the original SDSS spectra with 1 errors (York et al. 2005; York et al.
2006).

log Mg I Mg II Mg II Al III Al III Si II Ca II Ti II Zn IIa Zn IIb

HI 2852 2796 2803 1854 1862 1808 3933 3383 2026 2062

Q01230058 20.08 (799 95) (1911 75) (1895 78) 616 34 297 31 154 27 <7 <10
Q0132+0116 19.70 469 35 318 41 <7 <7
Q01380005 19.81 408 13 1186 15 1065 10 <40 <44 <75 (325 59) <7 131 31 <19
Q0153+0009 19.70 559 34 2563 22 2058 24 84 25 33 13 <27 (319 60) 46 11 <7 <11
Q04491645 20.98 287 16 346 9 270 15 257 20 33 6 160 15
Q2335+1501 19.70 432 10 974 11 1268 8 193 25 <7 81 25 <12
z abs Mn II Mn II Mn II Fe II Fe II Fe II Fe II Fe II Fe II Cr II
2576 2594 2606 2260 2344 2374 2382 2586 2600 2056
Q01230058 1.4094 74 13 118 34 1086 31 740 21 1247 27 <10
Q0132+0116 1.2712 <8 <6 <7 <7 1184 23 518 31 1744 20 1023 26 1782 20 <6
Q01380005 0.7821 <11 898 12 <20
Q0153+0009 0.7714 (580 87) (1258 77) (944 76) (1326 73) 21 5
Q04491645 1.0072 154 11 125 21 64 16 1156 10 803 11 1465 9 1174 12 1559 12 93 18

Downloaded from http://mnras.oxfordjournals.org/ by guest on September 8, 2016

Q2335+1501 0.6798 (295 84) 56 11 <11
a The contribution of Mg I 2026 is not taken into account in the estimate of the EW of Zn II 2026.
b The contribution of Cr II 2062 is not taken into account in the estimate of the EW of Zn II 2062.

of an emission line thus rendering continuum fitting non-trivial. where (g i)med is the median (g i) colour of a control sample
Ti II ( 3073, 3230, 3242, 3384) and Na I ( 3303, 3304) are of 100 quasars with the closest magnitude to the object within a
covered but not detected. Upper limits from non-detections lead to redshift interval of 0.05. The resulting error estimate is the colour
log N(Ti II) <11.36 and log N(Na I) < 12.75. The Mn II 2576, dispersion of this control sample combined with the error in the
2594, 2606 lines for that systems fall in a spectral gap. The results colour measure as done in Vladilo et al. (2006). (g i) colours
of the profile fitting analysis are summarized in Table 7, while Fig. 6 are calculated from SDSS DR5 release photometric values. The
illustrates the fits. object Q04491645 is not a SDSS quasar and therefore no (g
i) is available for this system. For Q0132+0116, SDSS DR3 release
photometric values were used, since the DR5 release values are not
3.3 Results: equivalent widths, total column densities and
available. The uncertainty in (g i) is large (see Table 10), so
abundances
that zero reddening systems could have values of (g i) between
For each transition, we determine the rest-frame equivalent width 0.2 and +0.2 (York et al. 2006).
(EW) value or limit at the expected position of the line. The 3
errors for the equivalent widths reflect both uncertainties in the con-
tinuum level and photon noise. If a certain line was not detected, the 4 DISCUSSION
limiting equivalent widths are 3 limits, based on a 3-pixel wide
line (after rebinning to 0.05 /pixel). Results from this study are 4.1 Metal-rich systems
shown in Table 8, numbers in parentheses are from the original In the sample presented here, we find two new solar or above so-
SDSS spectra with 1 errors (York et al. 2005; York et al. 2006). lar sub-DLAs towards Q01380005 with [Zn/H] = +0.28 0.16
When overlapping, the two are broadly in agreement. with rather high metallicity in iron too: [Fe/H] > 0.09 (i.e. Fe
Total column densities for all observed species are presented in could be depleted compared to Zn by a factor of 2 or more) and
Table 9 for each of the six absorption systems studied here. For each towards Q2335+1501 with [Zn/H] = +0.07 0.34. One of these
system, the top line presents the sum of the column densities ob- high-Zn abundance systems (Q01380005) has large (g i) =
tained from the Voigt profile fits while the column densities derived +0.45 0.13, and thus seems to be dusty, but not to have Fe sig-
from Apparent Optical Depth (AOD) method (Savage & Sembach nificantly depleted. The elements Si, Cr, Mn and Ti are depleted
1991) are shown on the line below. In all cases, both methods give compared to Zn. There could always be extinction arising in the
consistent results within the error estimates. We then estimate the circum quasar region which is hot and therefore does not show
abundances of elements with respect to solar for each system fol- absorption lines. Furthermore, there could be another absorption
lowing the standard definition: system at lower redshift that is causing the extinction. However, in
[X/Y] = log[N(X)/N(Y)]absorber log[N(X)/N(Y)] , (1) the case of Q01380005, we have checked that the SDSS spectrum
for this object shows no broad absorption line (BAL) feature and no
where log [N(X)/N(Y)] is the logarithmic solar abundance and is
identified narrow-line systems other than the target sub-DLA. We
taken from Asplund, Grevesse & Sauval (2005) adopting the mean
have checked that there are no Ca II at z < 0.7 and no Na I at z <
of photospheric and meteoritic values for Mg, Si, Mn, Fe, Zn, Ti, Co
0.4. The 1 EW limit of detection for Na I is 0.27 and for Ca II, it
and Ca. These values are recalled on the top line of Table 10. The
is 0.150.2 .
last column of the table summarizes the  (g i) for each absorber,
It is interesting to note that the situation with Fe and Si is reversed
where (g i) is defined as
from the case of the SMC, where some sightlines have Si undepleted
(g i) = (g i) (g i)med , (2) but Fe depleted (e.g. Welty, Hobbs & Morton 2003). Also, some of


C 2008 The Authors. Journal compilation 
C 2008 RAS, MNRAS 386, 22092220
 Metal abundances at z < 1.5 2217
Table 9. Logarithmic values of the total column densities for each absorber. For each entry, the first line refers to Voigt profile fitting and the second to the
Apparent Optical Depth (AOD) technique. Co is detected towards Q04491645 and Ca towards Q2335+1501. Transitions marked are not covered by our
spectra.

log N(H I) z abs Mg I Mg II Al III Si II Cr II

Q01230058 20.08+0.10
0.08 1.4094 13.76 0.02 15.47 0.05 12.91 0.10
AOD 13.67 0.05 15.50 0.07 12.88 0.15
Q0132+0116 19.70+0.08
0.10 1.2712 13.58 0.02 <12.13a
AOD 13.55 0.03
Q01380005 19.81+0.06
0.11 0.7821 12.66 0.01 >15.57 <12.28a <14.89a <12.61a
AOD 12.66 0.01 >14.29
Q0153+0009 19.70+0.08
0.10 0.7714 12.88 0.01 >15.76 12.60 0.XX <14.49a 12.81 0.10
AOD 12.84 0.02 >14.36 12.77 0.11 12.78 0.08
Q04491645 20.98+0.06
0.07 1.0072 12.37 0.01 13.58 0.02 15.86 0.04 13.47 0.02
AOD 12.40 0.02 13.61 0.02 15.73 0.03 13.44 0.08
Q2335+1501 19.70+0.30
0.30 0.6798 12.68 0.01 >15.45 12.89 0.10
AOD 12.67 0.01 >14.39 13.04 0.22
log N(H I) z abs Mn II Fe II Zn II Ti II Co II or Ca II Na I
Q01230058 20.08+0.10 1.4094 12.61 0.12 14.98 0.02 12.23 0.10

Downloaded from http://mnras.oxfordjournals.org/ by guest on September 8, 2016

0.08
AOD 12.62 0.06 15.08 0.12 12.41 0.19
Q0132+0116 19.70+0.08
0.10 1.2712 <11.42a 14.62 0.01 <11.76a
AOD 14.62 0.03
Q01380005 19.81+0.06
0.11 0.7821 <11.84a >15.17 12.69 0.05 <11.48a <13.05a
AOD >14.32 12.91 0.10
Q0153+0009 19.70+0.08
0.10 0.7714 <11.96a 12.02 0.06 <13.01a
AOD 12.13 0.09
Q04491645 20.98+0.06
0.07 1.0072 12.91 0.03 15.09 0.01 12.62 0.07 <11.47a 13.51 0.07 <12.66a
AOD 12.92 0.03 15.07 0.03 12.32 0.08 13.49 0.04
Q2335+1501 19.70+0.30
0.30 0.6798 14.83 0.03 12.37 0.04 <11.36a 12.41 0.03 <12.75a
14.74 0.08 12.69 0.13 12.40 0.05
a 3 limit calculated from the non-detection of the line.

Table 10. Abundances with respect to solar, [X/H]. The error bars on [X/H] include both the errors in log N(X) and log N(H I). Co is detected towards
Q04491645 and Ca towards Q2335+1501 (the abundance is derived from Ca II only). Transitions marked are not covered by the current spectra.

A(X/H) Mg Si Cr Mn Fe Zn Ti Co/Ca (g i)

4.47 4.46 6.37 6.53 4.55 7.40 7.11 7.11/5.69

Q01230058 0.150.15 0.800.20 0.940.21 0.550.12 0.450.20 +0.160.16

Q0132+0116 <1.20a <1.75a 0.530.12 <0.54a 0.070.12
Q01380005 >+0.23 <0.46a <0.83a <1.44a >0.09 +0.280.16 <1.22a +0.450.13
Q0153+0009 >+0.53 <0.75a 0.520.20 <0.34a 0.570.16 0.130.15
Q0449 1645b 0.660.11 1.140.09 1.540.11 1.340.08 0.960.08 <2.11a 0.360.14
Q2335+1501 >2.55 0.440.40 0.320.33 +0.070.34 <1.23 1.600.33 0.040.15
a 3 limit calculated from the non-detection of the line.
b Q04491645 was not observed by SDSS, therefore no (g i) could be calculated.

the dust might not be made of Fe, thus explaining the (g i) One more absorber may be dust free and have relatively high
reported here (but see also Vladilo 2002; Li, Misselt & Wang 2006; [Zn/H], Q01230058 with [Zn/H] = 0.45 0.20, while two sys-
Sofia et al. 2006). tems with upper limits on Zn could have higher [Zn/H] than most
The sub-DLAs in this study may need some ionization correc- DLAs, but new data are needed to confirm this: Q0132+0116 with
tions: as shown by Meiring et al. (2008), the corrections for stronger [Zn/H] < 0.54 and Q0153+0009 with [Zn/H] < 0.34 compared
ionizing radiation fields may lead to Fe depletion compared to Zn to the DLA (Q04491645 with [Zn/H] = 0.96 0.08) and to
and Cr, but should not affect Ti, which has an ionization potential DLAs in general (Kulkarni et al. 2007). These results are therefore
near that of H I. Indeed, simulation with the CLOUDY photoioniza- in line with our recent findings (Peroux et al. 2006a; Khare et al.
tion model shows that [Zn/Cr] is affected by ionization. Much of the 2007; Kulkarni et al. 2007) that sub-DLAs show higher metallicities
correction needed is Zn though, which is the most affected element at low redshift than classical DLAs. These systems may represent
of the two. But this could be partly due to uncertain atomic data for the tip of the iceberg of a metal-rich population of higher H I column
Zn, and in any case ionization corrections for Zn lead to higher Zn density systems currently hidden by dust extinction effects. If there
abundances and higher [Zn/Cr] (e.g. Vladilo et al. 2001; Meiring really are very few DLAs with solar metallicity, then DLAs may
et al. 2007). mainly come from dwarf galaxies (Khare et al. 2007). In any case,


C 2008 The Authors. Journal compilation 
C 2008 RAS, MNRAS 386, 22092220
2218 C. Peroux et al.
the fact that the average metallicity in observed DLAs is lower than
in observed sub-DLAs would be partly explained.

4.3 Abundances
Calcium II: Ca II H and K lines might constitute a secure tool with
which to detect DLA at low redshift even from medium-resolution
optical spectra. Wild, Hewett & Pettini (2006) show that the 37
Ca II absorbers found in SDSS, are remarkably dusty. They also
report a clear correlation between the strength of the Ca II equiva-
lent width and the amount of reddening in the background quasar
spectra. These findings are surprising given that Ca II is known to
be highly depleted on to dust grains and that, therefore, one would
expect dusty absorbers to have small measured gas phase Ca II abun-
dances. To better understand the relation between observed Ca II and
dust, higher resolution data of quasar absorbers with N(H I) mea-
sured where the dust content can be estimated with other means
Figure 7. (g i) versus [Cr/Zn] for sub-DLAs and DLAs from this study
and from the literature (Meiring et al. 2007, 2008; Prochaska et al. 2006;
(i.e. Cr II/Zn II) are required. Here, we report a new high-resolution
measure of Ca II column density in an absorber at z abs = 0.6798 to-

Downloaded from http://mnras.oxfordjournals.org/ by guest on September 8, 2016

Peroux et al. 2006a,b; Vladilo et al. 2006). The stars are measurements
or upper/lower limits (as indicated by the arrows). There is still too few wards Q2335+1501. We derive log N(Ca II) =12.41 0.03. Given
measurements to show a strong relation between these ratios and the amount that some Ca III might also be present in H I region, this leads to an
of reddening measured by (g i). upper limit: [Zn/Ca] 1.67 (while [Zn/Cr] = +0.54 and [Zn/Fe] =
+0.39). However, Ca is not be a reliable indicator of the presence
or absence of dust in the disc. Indeed, if dust is present anywhere
the high abundance sub-DLAs are comparable with the ones mea- in the disc, the RoutlySpitzer effect (Routly & Spitzer 1952; see
sured in actively star-forming galaxies such as the Lyman-Break below) could still produce high [Ca/H] in some of the gas.
Galaxies at higher redshifts (Pettini et al. 2002), and therefore offer Cobalt II: Almost no Co is found in the ISM, so the reference
a crucial bridge between galaxies studied in emission and galaxies standard is tough to set. Based on the results of Mullman et al.
studied in absorption. (1998), Co may be somewhat less depleted in the Galaxy than Fe
(perhaps similar to Cr in cold clouds, slightly less depleted in warm
clouds). Ellison, Ryan & Prochaska (2001b) reported the first de-
4.2 Dust content
tection of cobalt in a quasar absorber with [Co/H] = 0.50 and
As opposed to [Zn/H], most of the [Fe/H] measurements presented [Co/Fe] = +0.31 0.05. Since then, other detections have been
here are subsolar (only Q01380005 has [Fe/H] > 0.09, i.e. not reported by Rao et al. (2005) with [Co/Fe] = +0.05 0.12 and
definitely subsolar). The ratio of depleted (e.g. Cr, Fe) to undepleted Dessauges-Zavadsky et al. (2006) with [Co/Fe] = +0.24 0.13.
(Zn) elements provides a measure of the dust content of quasar ab- Here, we present a new high-resolution detection of cobalt in the
sorbers (e.g. Pettini et al. 1997). The larger the deficiency relative DLA towards Q04491645 with log N(Co II) =13.51 0.07 lead-
to the solar ratio [Cr/Zn] = 0, the larger is the fraction of chromium ing to [Co/H] = 0.36 0.14 and [Co/Fe] = +0.98. This value is
depleted from the gas to the dust phase. The refractory elements quite high, but it is not excluded that an interloper is responsible for
such as Cr and Fe might be expected to show a correlation be- the absorption instead of Co II.
tween their abundances relative to the mildly depleted elements Titanium II: The few abundance measures of Ti to date in DLAs
such as Zn and the inferred reddening. We find [Cr/Zn] = 0.35 in suggest that the relative abundance of Ti with respect to Fe is per-
Q01230058, [Cr/Zn] < 0.42 in Q01380005, [Cr/Zn] > 0.18 haps slightly supersolar. Since stellar Ti abundances behave as an
in Q0153+0009 and [Cr/Zn] = 0.51 in Q2335+1501. Finally, -capture element (such as Mg, Si and S) in the Galaxy, this may in-
[Cr/Zn] = 0.18 in the DLA towards Q04491645. There are still dicate both low-dust depletion and/or be evidence for an -product
too few measurements to show a strong relation between these ra- enhancement for Type II supernovae (SNe). Ti is normally a highly
tios and the amount of reddening measured by (g i). Indeed, as depleted element, so high [Ti/Zn] means low dust. Furthermore,
shown in Fig. 7, we observe large depletions as measured by [Cr/Zn] the interpretation is mitigated by the RoutlySpitzer effect where
being associated with both positive (Q01230058) and negative Fe in dust grains might be released in the high-velocity clouds and
(Q2335+1501) values of (g i). In this redshift regime, E(B strengthen the wings of Fe II relative to Mg II (York et al. 2006).
V) 0.25 (g i): the extinction in the three systems with Thus far, only a few Ti measurements exist in DLAs at z abs < 1.5
the lowest (g i) may be <0.05 in E(B V) but the depletions (Dessauges-Zavadsky et al. 2002; Ledoux et al. 2002), while ef-
can still be significant (e.g. York & Kinahan 1979). The only DLA forts to measure Ti in DLAs at z abs > 1.5 have focused on the pair
in our sample towards Q04491645 seems to have low extinction, of weak lines at rest 1910 and have yielded only a few ten-
consistent with observed trends for zinc in DLAs. This is consistent uous detections (e.g. Prochaska & Wolfe 1997, 1999; Prochaska,
with findings that DLAs show little depletion while sub-DLAs show Gawiser & Wolfe 2001). Our observations of the stronger Ti II
considerable depletion at times (e.g. table 12 of Meiring et al. 2007; 3073, 3242, 3384 lines have resulted in one robust detection and
table 14 of Meiring et al. 2008). In fact, if dust extinction in quasar two significant upper limits. We derive [Ti/H] = 0.57 0.16 in
absorbers is a strong function of N(Zn II), the obscuration will start Q0153+0009; [Ti/H] <1.22 in Q01380005; [Ti/H] <2.11 in
acting at a lower [Zn/H] ratio for DLAs than for sub-DLAs (Vladilo Q04491645 and [Ti/H] <1.23 in Q2335+1501. The range ra-
& Peroux 2005). In other words, we might be missing relatively tios N(Ti II)/N(H I) is consistent with the range in diffuse interstellar
more systems in the DLA range than in the sub-DLA range. Then, clouds: the depletion is as much as a factor of 100, in our case.


C 2008 The Authors. Journal compilation 
C 2008 RAS, MNRAS 386, 22092220
 Metal abundances at z < 1.5 2219
Sodium I: Finally, we note that while lines of Na I were covered in that the abundance variation depends on uncertain ionization cor-
several lines of sight, we did not detect the 3302 doublet absorption rections, current photoionization models (Meiring et al. 2007, 2008)
in any system. This is not surprising, given the general weakness of do suggest the variations are in abundance, not radiation fields being
these features in low H I column lines of sight in the Galaxy. different from component to component.

4.4 On the velocity-spreadmetallicity relation AC K N OW L E D G M E N T S

In three of the six absorbers studied here, we could determine the We thank Sandhya Rao for providing N(H I) estimate of the absorber
equivalent width of Mg II 2796 from the high-resolution UVES towards Q2335+1501 in advance of publication. We would like to
data. For these systems, we can thus derive the D-indices defined thank the Paranal and Garching staff at ESO for performing the ob-
by Ellison (2006), where, as in the original work,  v is the total servations in service mode. VPK and JM acknowledge partial sup-
line width excluding detached absorption components where the port from the US National Science Foundation grant AST-0607739
continuum is recovered and only including the complex profile and (PI: Kulkarni).
the EW is expressed in :
REFERENCES
D = 1000 EW(Mg II2796)/v. (3)
This gives D index = 5.9, 5.6 and 5.4 for absorbers towards Asplund M., Grevesse N., Sauval A. J., 2005, in Barnes T. G., III, Bash F. N.,
Q01380005, Q0153+0009 and Q2335+1501, respectively. So eds, ASP Conf. Ser. Vol. 336, Cosmic Abundances as Records of Stellar

Downloaded from http://mnras.oxfordjournals.org/ by guest on September 8, 2016

these three sub-DLAs, indeed have a D-index below the 6.3 limit Evolution and Nucleosynthesis. Astron. Soc. Pac., San Francisco, p. 25
Bergeron J., DOdorico S., 1986, MNRAS, 220, 833
indicating log N(H I) < 20.3.
Blades J. C., Hunstead R. W., Murdoch H. S., Pettini M., 1985, ApJ, 288,
On the other hand, it has been recently advocated that the well- 580
known galaxy luminositymetallicity relation observed at 0 < z < Bolton J. G., Wall J. V., 1969, ApJ, 3, L177
1 (Tremonti et al. 2004) was already in place at higher redshifts. Bouche N., Lehnert M., Peroux C., 2005, MNRAS, 364, 319
In fact, a massmetallicity relation has recently been put into evi- Bouche N., Lehnert M., Peroux C., 2006, MNRAS, 367, 16
dence for UV-selected star-forming galaxies at z 2.3 by Erb et al. Bouche N., Murphy M., Peroux C., Davies R., Eisenhauer F., Forster
(2006). An empirical relation between velocity spread and metal- Schreiber N., Tacconi L., 2007, ApJ, 669, L5
licity in quasar absorbers was independently reported by Peroux Dessauges-Zavadsky M., Prochaska J. X., DOdorico S., 2002, A&A, 391,
et al. (2003b). More recently, Ledoux et al. (2006) have used this 801
relation to claim a massmetallicity relation in high-redshift ab- Dessauges-Zavadsky M., Prochaska J. X., DOdorico S., Calura F.,
Matteucci F., 2006, A&A, 445, 93
sorbers, assuming that the velocity width of quasar absorbers is
Ellison S., 2006, MNRAS, 368, 335
a good proxy for halo masses. Meiring et al. (2007) also com- Ellison S., Yan L., Hook I. M., Pettini M., Wall J. V., Shaver P., 2001a, A&A,
pared the metallicity-velocity width correlation for sub-DLAs and 379, 393
DLAs, as did Prochaska et al. (2008) for the Gamma Ray Burst host Ellison S., Ryan S. G., Prochaska J. X., 2001b, MNRAS, 326, 628
galaxies. Erb D. K., Shapley A. E., Pettini M., Steidel C. C., Reddy N. A., Adelberger
We note that this empirical correlation is expected to relate to K. L., 2006, ApJ, 644, 813
the evolution of the D-index with log N(H I). However, the interpre- Ferrara A., Scannapieco B. J., 2005, ApJ, 634, L37
tation of  v as a measure of the halo mass (Haehnelt, Steinmetz Fox A., Petitjean P., Ledoux C., Srianand R., 2007, A&A, 465, 171
& Rauch 1998; Maller et al. 2001) is challenged by observational Haehnelt M. G., Steinmetz M., Rauch M., 1998, ApJ, 495, 647
evidences. In fact, recent H I-gas rich observations of local galax- Jenkins E. B., Bowen D. V., Tripp T. M., Sembach K. R., 2005, ApJ, 623,
767
ies by Zwaan et al. (submitted), show that the velocity spread is
Khare P., Kulkarni V. P., Lauroesch J. T., York D. G., Crotts A. P. S.,
not a good indicator of mass in H I-gas rich DLA analogues at z = Nakamura O., 2004, ApJ, 616, 86
0. Similarly, Bouche et al. (2007) have recently suggested that the Khare P., Kulkarni V. P., Peroux C., York D. G., Lauroesch J. T., Meiring J.
equivalent width of Mg II absorbers is due to winds, not gravity D., 2007, A&A, 464, 487
related velocity dispersion. Kulkarni V. P., Fall S. M., 2002, ApJ, 580, 732
Kulkarni V. P., Fall S. M., Lauroesch J. T., York D. G., Welty D. E., Khare
P., Truran J. W., 2005, ApJ, 618, 68
5 CONCLUSIONS Kulkarni V. P., Khare P., Peroux C., York D. G., Lauroesch J. T., Meiring J.
D., 2007, ApJ, 661, 88
We have used new high-resolution observations of six quasar ab- Ledoux C., Bergeron J., Petitjean P., 2002, A&A, 385, 802
sorbers to constrain the metallicity of sub-DLAs in the low-redshift Ledoux C., Petitjean P., Fynbo J. P. U., Moller P., Srianand R., 2006, A&A,
Universe. We specially targeted the Zn II line which is expected to be 457, 71
a metallicity indicator free from the bias effect of dust. In line with Li A., Misselt K. A., Wang Y. J., 2006, ApJ, 640, L151
recent results from our group (Peroux et al. 2006a,b; Meiring et al. Malaney R. A., Chaboyer B., 1996, ApJ, 462, 57
2007; Khare et al. 2007; Kulkarni et al. 2007; Meiring et al. 2008), Maller A. H., Prochaska J. X., Somerville R. S., Primack J. R., 2001, MN-
we find that sub-DLAs appear to be more metal-rich than classical RAS, 326, 1475
Meiring J. D. et al., 2006, MNRAS, 370, 43
DLAs. In particular, we report the discovery of two sub-DLAs with
Meiring J. D., Lauroesch J. T., Kulkarni V. P., Peroux C., Khare P., York D.
supersolar metallicity towards Q01380005 with [Zn/H] = +0.28
G., Crotts A. P. S., 2007, MNRAS, 376, 557
0.16 and Q2335+1501 with [Zn/H] = +0.07 0.34. We note Meiring J. D., Kulkarni V. P., Lauroesch J. T., Peroux C., Khare P., York D.
that there are significant variations in column density ratios of met- G., Crotts A. P. S., 2008, MNRAS, 384, 1015
als from component to component in some of our systems, and also Morton D. C., 2003, ApJS, 149, 205
that there is some indication of different grain types being required Mullman K. L., Lawler J. E., Zsargo J., Federman S. R., 1998, ApJ, 500,
if the depletion is related to extinction. While it is still possible 1064


C 2008 The Authors. Journal compilation 
C 2008 RAS, MNRAS 386, 22092220
2220 C. Peroux et al.
Pei Y. C., Fall S. M., Hauser M. G., 1999, ApJ, 522, 604 Rao S. M., Prochaska J. X., Howk J. C., Wolfe A. M., 2005, AJ, 129, 9
Petitjean P., 1998, in Le Fevre O., Charlot S., eds, Les Houches school Rao S., Turnshek D., Nestor D., 2006, ApJ, 636, 610
Formation and Evolution of galaxies. Springer-Verlag, Berlin Routly P. M., Spitzer L. Jr, 1952, ApJ, 115, 227
Peroux C., McMahon R. G., Storrie-Lombardi L. J., Irwin M., 2003a, Savage B. D., Sembach K. R., 1991, ApJ, 379, 245
MNRAS, 346, 1103 Sofia U. J., Gordon K. D., Clayton G. C., Misselt K., Wolff M. J., Cox N. L.
Peroux C., Dessauges-Zavadsky M., DOdorico S., Kim T. S., McMahon J., Ehrenfreund P., 2006, ApJ, 636, 753
R., 2003b, MNRAS, 345, 480 Sommer-Larsen J., Fynbo J., 2008, MNRAS, in press (astro-ph/0703698)
Peroux C., Dessauges-Zavadsky M., DOdorico S., Kim T. S., McMahon R. Tremonti C. A. et al., 2004, ApJ, 613, 898
G., 2005, MNRAS, 363, 479 Vidal-Madjar A., Laurent C., Bonnet R. M., York D. G., 1977, ApJ, 211,
Peroux C., Kulkarni V. P., Meiring J., Ferlet R., Khare P., Lauroesch J. T., 91
Vladilo G., York D. G., 2006a, A&A, 450, 53 Vladilo G., 2002, ApJ, 569, 295
Peroux C., Meiring J., Kulkarni V. P., Ferlet R., Khare P., Lauroesch J. T., Vladilo G., 2004, A&A, 421, 479
Vladilo G., York D. G., 2006b, A&A, 373, 369 Vladilo G., Peroux C., 2005, A&A, 444, 461
Pettini M., 2006, in LeBrun V., Mazure A., Arnouts S., Burgarella D., eds, Vladilo G., Centurion M., Bonifacio P., Howk J. C., 2001, A&A, 557, 1007
The Fabulous Destiny of Galaxies: Bridging Past and Present. Frontier Vladilo G., Centurion M., Levshakov S. A., Peroux C., Khare P., Kulkarni
Group, Paris, p. 319 V. P., York D. G., 2006, A&A, 454, 151
Pettini M., Hunstead R. W., Murdoch H. S., Blades J. C., 1983, ApJ, 273, Welty D. E., Hobbs L. M., York D. G., 1991, ApJS, 75, 245
436 Welty D. E., Hobbs L. M., Morton D. C., 2003, ApJS, 147, 61
Pettini M., Smith L. J., King D. L., Hunstead R. W., 1997, ApJ, 486, 665 Wild V., Hewett P., Pettini M., 2006, MNRAS, 367, 211
Pettini M., Ellison S. L., Steidel C. C., Shapley A. E., Bowen D. V., 2000, York D. G. et al., 2005, in Williams P. R., Shu C.-G., Menard B., eds, Proc.

Downloaded from http://mnras.oxfordjournals.org/ by guest on September 8, 2016

ApJ, 532, 65 IAU Colloq. 199, Probing Galaxies through Quasar Absorption Lines.
Pettini M., Rix S. A., Steidel C. C., Hunt M. P., Shapley A. E., Adelberger Cambridge Univ. Press, Cambridge, p. 58
K. L., 2002, Ap&SS, 281, 461 York D. G., Kinahan B. F., 1979, ApJ, 228, 127
Prochaska J. X., Wolfe A. M., 1997, ApJ, 474, 140 York D. G. et al., 2006, MNRAS, 367, 945
Prochaska J. X., Wolfe A. M., 1999, ApJS, 121, 369 Zwaan M., Walter F., Brinks E., De Block W. J. G., Kennicutt R. C.,
Prochaska J. X., Gawiser, Eric, Wolfe A. M., 2001, ApJ, 552, 99 Ryan-Weber E., 2008, ApJ, submitted
Prochaska J. X., Chen H.-W., Wolfe A. M., Dessauges-Zavadsky M., Bloom
J. S., 2008, ApJ, 672, 59
Prochaska J. X., Herbert-Fort S., Wolfe A. M., 2005, ApJ, 635, 123
Prochaska J. X., OMeara J. M., Herbert-Fort S., Burles S., Prochter G. E.,
Bernstein R. A., 2006, ApJ, 648, 97 This paper has been typeset from a TEX/LATEX file prepared by the author.


C 2008 The Authors. Journal compilation 
C 2008 RAS, MNRAS 386, 22092220