EDP Sciences logo
Subscriber Authentication Point
Search
Menu
Home All issues Volume 594 (October 2016) A&A, 594 (2016) A91 Full HTML
Free Access
Issue
A&A
Volume 594, October 2016
Article Number A91
Number of page(s) 26
Section Catalogs and data
DOI https://doi.org/10.1051/0004-6361/201628161
Published online 18 October 2016

© ESO, 2016

1. Introduction

In the era of massive quasar (QSO) surveys, already encompassing hundreds of thousands of confirmed sources (e.g., Pâris et al. 2014; Flesch 2015), there is a relative shortage of follow-up echelle quality spectroscopy. Moderate to high resolving power (R ≈ 5000−40 000) and wide spectral coverage are key to many absorption line diagnostics that probe the interplay between galaxies and the intergalactic medium (IGM) at all redshifts. However, such observations are time consuming and require large telescopes, and even more so for high redshift QSOs which tend to be faint. Another challenge for QSO absorption line science is that as the redshift increases, more of the rest-frame UV and optical transitions become shifted into the hard-going near-infrared (NIR; 1 μm ≲ λ ≲ 2.5 μm). Presently, public archives contain echelle spectra of roughly a few thousand unique QSOs, of which just a small fraction has NIR coverage. In addition, these data arise primarily from the cumulative effort of single (and heterogenous) observing programs, so one would expect such databases to be inhomogeneous in nature and suffer from selection biases by construction (Brunner et al. 2002; Djorgovski 2005). Thus, new homogeneous and statistically significant echelle data sets are always welcome with as wide a range of uses as possible. In this paper we present “XQ-100”, a new legacy survey of 100 QSOs at emission redshifts zem ≃ 3.5−4.5 observed with full optical and NIR coverage using the echelle spectrograph X-shooter (Vernet et al. 2011) on the European Southern Observatory (ESO) Very Large Telescope (VLT). The context and the scientific motivation of the survey are as follows.

The largest QSO echelle samples in the optical come from Keck/HIRES (“KODIAQ” database; O’Meara et al. 2015) and VLT/UVES (ESO UVES public archive) each providing between 300 and 400 QSO spectra with R ≈ 40 000. At moderate resolving power, R ≈ 10 000, Keck/ESI has observed around a thousand QSOs (John O’Meara, private communication) and a search in the VLT/X-shooter public archive reveals spectra of almost 300 sources to date. Other large optical facilities with echelle capabilities, such as Subaru or Magellan, have either acquired a smaller data volume or do not manage public archives. In addition to “smaller” programs (≲ 10 targets), these data sets, public or not, have been fed over the years by a few dedicated QSO surveys (e.g., Bergeron et al. 2004) aimed at a variety of astrophysical probes of galaxy evolution and cosmology: metals in damped Lyα systems (DLAs; e.g., Lu et al. 1996; Prochaska et al. 2003; Ledoux et al. 2003; Rafelski et al. 2013) and in the IGM (e.g., Aguirre et al. 2004; Songaila 2005; Scannapieco et al. 2006; D’Odorico et al. 2010); light elements in Lyman-limit Systems (e.g., Kirkman et al. 2003); DLA galaxies (e.g., Peroux et al. 2011; Noterdaeme et al. 2012a; Zafar et al. 2013); low and high-z circum-galactic medium (e.g., Chen et al. 2010; Rudie et al. 2012); thermal state of the IGM (e.g., Schaye et al. 2000; Kim et al. 2002); reionization (e.g., Becker et al. 2007, 2012, 2015); matter power spectrum (e.g., Croft et al. 2002; Viel et al. 2004, 2009, 2013); and fundamental constants (e.g., Murphy, et al. 2003; Srianand et al. 2004; Molaro et al. 2013).

In the NIR, the largest QSO spectroscopic survey so far has been conducted using the FIRE IR spectrograph at Magellan (Matejek & Simcoe 2012). Focused on the incidence of Mg ii at z ≈ 2–5, this survey comprises NIR observations of around 50 high-z QSOs at R ≈ 6000 and median signal-to-noise ratio, S/N = 13. Other surveys at moderate to high resolution have focused on the C iv mass density at z> 4 using Magellan/FIRE (Simcoe et al. 2011), Keck/NIRSPEC (Becker et al. 2009; Ryan-Weber et al. 2009; Becker et al. 2012), or VLT/X-shooter (D’Odorico et al. 2013), albeit comprising only a handful of sightlines, given the paucity of very high-z QSOs.

Near-IR spectroscopy is also needed to study the rest-frame optical emission lines of high-z QSOs, which constrain broad-line region metallicities and black hole masses; however, in this case spectral coverage is more important than resolution. For instance, surveys have used VLT/ISAAC (Sulentic et al. 2006, 2004; Marziani et al. 2009), NTT/SofI (Dietrich et al. 2002, 2009), or Keck/NIRSPEC and Blanco/OSIRIS (Dietrich et al. 2003). There are also samples at higher resolution obtained with Gemini/GNIRS (Jiang et al. 2007), or VLT/X-shooter (Ho et al. 2012; De Rosa et al. 2014). The largest samples have been acquired using Palomar Hale 200-inch/TripleSpec (Zuo et al. 2015, 32 QSOs at 3.2 <z< 3.9) and, at lower redshifts, VLT/X-shooter (Capellupo et al. 2015, 30 QSOs at z ≈ 1.5).

The present XQ-100 survey builds on observations made with VLT/X-shooter within the ESO Large Programme entitled “Quasars and their absorption lines: a legacy survey of the high redshift Universe with X-shooter” (PI S. López; 100 h of Chilean time). X-shooter provides complete coverage from the atmospheric cutoff to the NIR in one integration at R ≈ 6000–9000, depending on wavelength. The full spectral coverage, along with a well-defined target selection and the high S/N achieved (median S/N = 30), clearly make XQ-100 a unique data set to study the rest-frame UV/optical of high-z QSOs in a single, homogeneous, and statistically significant sample. Our program was based on the following scientific themes:

Galaxies in absorption: determining the cosmic density of neutral gas in DLAs, the main reservoirs of neutral gas in the Universe (e.g., Wolfe et al. 2005; Prochaska & Wolfe 2009; Noterdaeme et al. 2012b) at z> 3.5 (Sánchez-Ramírez et al. 2016); studying individual DLA abundances at 2.0 ≲ z ≲ 4.5 (Berg et al. 2016); constraining the Mg ii incidence (dN/ dz)Mg ii at z> 2.5 with ~ 2–3 times better sensitivity and ~ 2 times longer redshift path than the sample by Matejek & Simcoe (2012) to test predictions from the cosmic star formation rate (Zhu & Ménard 2013; Ménard et al. 2011).

Intergalactic-Medium science: measuring the cosmic opacity at the Lyman limit (Prochaska et al. 2009; Worseck et al. 2014) and providing an independent census of Lyman-limit systems (LLS; Prochaska, O’Meara & Worseck 2010; Songaila & Cowie. 2010) at z ≃ 1.5−4.5; constraining the UV background via the proximity effect (e.g., D’Odorico et al. 2008; Dall’Aglio et al. 2008; Calverley et al. 2011).

Active-Galactic-Nuclei science: making the first z> 3.5 accurate measurements of black hole masses using the rest-frame UV emission lines of C ivλ1549 and Mg iiλ2800 and the rest-frame optical Hβ line (from line widths and continuum luminosities; e.g., Vestergaard & Peterson 2006; Vestergaard & Osmer 2009); examining broad-line region metallicity estimates (from emission line ratios; e.g., Hamann & Ferland 1999; Hamann et al. 2002) and their relationship with other QSO properties, including, but not limited to, luminosity and black hole mass; using associated absorption lines to study the co-evolution of galaxies and black holes by measuring metallicities in the interstellar-medium of the QSO host galaxies (Perrotta et al. 2016; D’Odorico et al. 2004); studying the broad QSO-driven outflow absorption lines that are found serendipitously in the spectra.

Cosmology: measuring the matter power spectrum with the Lyα forest (Croft et al. 1998) at high redshift (e.g., Viel et al. 2009; Palanque-Delabrouille et al. 2013), including an independent measurement of cosmological parameters with a joint analysis of these and the Planck publicly released data (Iršič et al. 2016).

The sample size of 100 QSOs was defined by the objectives of these science goals. The choice of emission redshifts was determined by the absorption line searches: z ≳ 3.5 means that every QSO contributes a redshift path of at least 0.5 for (dN/ dz)Mg ii in the NIR, while z ≲ 4.5 avoids excessive line crowding in the Lyα forest. Clearly, a combination of the factors: well-defined target selection, echelle resolution, high S/N, and full wavelength coverage all represent a benefit to the above science goals.

XQ-100 was designed as a legacy survey and this paper accompanies the public release of all data products, including a uniform sample of 100 reduced X-shooter spectra1. We note that this data volume increases the X-shooter QSO archive by ≈ 30%.

The following sections provide an in-depth description of the survey, along with its basic statistics. A description of our target selection and the observations can be found in Section Sect. 2; details of the data reduction, along with a comparison between our own custom pipeline and the one provided by ESO are given in Sect. 3; details of data post-processing (telluric corrections and continuum fits) are given in Sect. 4; and, finally, a description of the publicly released data products is given in Sect. 5. For a technical description of the instrument, we refer the reader to Vernet et al. (2011) and to the online X-shooter documentation23.

2. Target selection and observations

2.1. Target selection

XQ-100 targets were selected initially from the NASA/IPAC Extragalactic Database (NED) to have emission redshifts z> 3.5 and declinations δ< + 15 degrees. To fill some right-ascension gaps lacking bright z> 3.5 targets, twelve additional targets with + 15 <δ< + 30 were selected from literature sources. Then the Sloan Digital Sky Survey Data Release 7 database (SDSS DR7; Schneider et al. 2010) was screened with the further criterion of having SDSS magnitude r< + 20. Finally, these candidates were cross-correlated with the Automate Plate Machine (APM) catalog4 to obtain uniform magnitudes in a single pass-band (R), which we also use throughout the present paper. Our primary selection is thus biased toward bright sources; however, as explained below, we made our best effort to minimize biases affecting the absorption line statistics.

We avoided targets with known broad absorption line features, and targets with an intrinsic color selection bias from the SDSS. The SDSS color selection is biased at the lower redshift end of our survey (z< 3.6, see Worseck & Prochaska 2011). Here, we required SDSS QSOs to be radio-selected or previously discovered with other techniques such as slitless spectroscopy. Without these precautions, our goal of obtaining a truly blind and unbiased target selection would have been undermined, despite the relatively small number of targets impacted. For example the SDSS color bias would result in (1) underestimates of the mean free path (Prochaska et al. 2009); (2) overestimates of the DLA –and also the LLS– incidence (Prochaska, O’Meara & Worseck 2010); (3) a higher metal dN/ dz due to the higher incidence of LLSs and partial LLSs; (4) a higher fraction of proximate LLSs that affect proximity effect studies; and (5) potentially a slight bias in the mean QSO spectral energy distribution towards red QSOs (Worseck & Prochaska 2011). We should also note that although earlier color survey designs (Palomar Spectroscopic Survey, APM BR, APM BRI) considered color selection effects at the low-z end (Irwin et al. 1991; Storrie-Lombardi et al. 1994), these were never well quantified. Thus, follow-up on color-selected QSOs close to the stellar locus should be done with care (or avoided altogether), as the sightlines are potentially biased in their LLS statistics.

During program execution we replaced four targets in our original list that had been observed by Matejek & Simcoe (2012) using Magellan/FIRE; however, we intentionally observed three other FIRE targets in order to have a reference in characterizing absorption line detection limits: J1020+0922 at z = 3.640, J1110+0244 z = 4.146, and J1621-0042 at z = 3.711.

Our final sample, taking into account the various selections described above and also considering the relative paucity of high redshift QSOs, has emission redshifts ranging from 3.508 to 4.716. Since the most distant QSO in our sample is the only target with zem> 4.5, for simplicity we refer to the redshift range of the survey as zem ≃ 3.5–4.5 throughout this paper.

thumbnail Fig. 1

Sky distribution of XQ-100 sources. The color scale indicates emission redshifts.

thumbnail Fig. 2

XQ-100 emission redshifts.

Figure 1 shows the sky distribution of the observed XQ-100 sample. A color scale depicts emission redshifts. Figures 2 and 3 show the final distribution of QSO emission redshifts and R-magnitudes, respectively.

The full target list is provided in Table A.1, along with basic target properties (see Sect. 3). A full catalog with all observed target properties (listed in Table A.2) is provided online along with the data5.

thumbnail Fig. 3

XQ-100 R-magnitudes (APM).

2.2. Observations

The observations were carried out in “service mode” between April 1, 2012, and March 26, 2014. During this time X-shooter was mounted on unit 2 of the VLT. Service mode allows the user to define the Observation Blocks (OBs), which contain the instrument setup and are carried out by the observatory under the required weather conditions.

Table 1 summarizes the requested conditions of XQ-100. The airmass constraint was set according to each target’s declination such that the target was observable above the set constraint for at least 2 h. The requested constraints on sky brightness were fraction of lunar illumination < 0.5 and minimum moon distance 45 degrees. The targets were split into two samples, brighter and fainter than magnitude RAPM = 18.0. The seeing constraint was set to 1.0′′ for the bright sample and 0.8′′ for the faint sample. ESO Large Programmes are granted high priority status, which means that observations out of specifications are repeated and eventually carried over to the following semester until the constraints are met (to within ≈ 10%). In our case 13 targets were observed more than once because of interrupted OBs or because of ADC issues (Sect. 2.2.1)6. As a consequence of this process, 88 XQ-100 targets were observed within specifications, and 12 almost within specifications (i.e., the constraints were worse by ≲ 10%).

Table 1

Requested observing conditions.

Table 2

Instrument setup.

Table 2 summarizes the instrument setup. X-shooter has three spectroscopic arms, UVB, VIS and NIR, each with its own set of shutter, slit mask, cross-dispersive element, and detector. In order to obtain signal-to-noise ratios, that are as uniform as possible, XQ-100 integration times varied across the samples and also across the three spectroscopic arms. The bright sample had two integrations, each with Texp = 890 s in UVB, Texp = 840 s in VIS and Texp = 900 s in the NIR. The faint sample had four exposures, each with Texp = 880 s in the UVB, Texp = 830 s in the VIS, and Texp = 900 s in the NIR. These conditions defined two classes of OBs, which – including acquisition – had a total of 39 and 70 min duration, respectively. In order to optimize the sky-subtraction in the NIR, the exposures were nodded along the slit by ± 2.5′′ from the slit center.

The adopted slit widths were 1.0′′ in the UVB and 0.9′′ in the VIS and NIR, to match the requested seeing and account for its wavelength dependence. These slit widths provide a nominal resolving power of 4350, 7450, and 5300, respectively. The slit position was always set along the parallactic angle, except for five targets for which it was necessary to avoid contamination of a nearby bright object in the slit; these cases are relevant to a problem with the atmospheric dispersion corrector system (see next Section). Target acquisition was done in the R filter. The UVB and VIS were binned by a factor 2 in the dispersion direction.

For emission redshifts z> 4, the [OIII]λ5007 emission line lies out of the K-band. For 4.0 ≲ z ≲ 4.5, [OII]λ3727 falls in the gap between the H- and K-bands. Therefore, the 53 XQ-100 sources having z> 4 were observed using a K-band blocking filter that lowers the sky background where scattered light from the K-band affects primarily the J-band (Vernet et al. 2011). No blocking filter was used for z< 4 sources (47) in order to include [OIII]λ5007 in the wavelength range. We note that Mg iiλλ2796, 2803 is always in the wavelength range. See Fig. 7 for an example of a spectrum presenting the above-mentioned emission lines.

For each exposure, the standard calibration plan of the observatory was used to observe a hot star for telluric corrections. This plan foresees the observation of a telluric standard within 2 h and 0.2 airmasses of each science observation (but see Sect. 4.1).

2.2.1. ADC issues

In March 2012 ESO reported that the atmospheric dispersion correctors (ADCs) of the UVB and VIS arms started to fail occasionally, leading to possible wavelength-dependent slit losses, potentially worse than if no ADCs were used. In August 2012 the ADCs were disabled for the rest of the observations (at the time of writing the causes of these failures are being investigated).

By August 2012, around 30% of the XQ-100 observations had been executed. After checking our spectra carefully, we noticed the ADC problem had possibly affected 12 of the spectra, which showed an unusually large flux mismatch between the arms (see example in top panel of Fig. 4, which is explained below). The reason for such a mismatch was probably that these targets had been observed at a high enough airmass for a malfunctioning ADC to lead to strong chromatic slit losses.

thumbnail Fig. 4

XQ-100 spectra of the same QSO, J1126−0124, taken with the faulty ADCs in April 2012 (top panel) and repeated with the disabled ADCs in February 2014 (middle panel). Both observations were executed at a similar airmass of ≈ 1.3 at the parallactic angle. The dashed lines indicate the boundaries of the X-shooter arms. The match is better between the VIS and NIR arms in the middle panel. The bottom panel shows the same February 2014 XQ-100 spectrum but smoothed and rebinned to SDSS resolution (blue line), and rescaled by a factor of 1.3 to match the corresponding SDSS spectrum (overlaid in red). The good match suggests that slit losses in the XQ-100 data are roughly achromatic.

Five out of these 12 OBs were executed a second time with the disabled ADCs and using the parallactic angle. The improvement was evident. The two upper panels of Fig. 4 show XQ-100 spectra of the same OB executed before and after the ADC disabling. We note the effect of the faulty ADCs on the flux levels and slope in the UVB and VIS arms only (top panel), while the NIR arm is not affected, which is expected since this arm does not use an ADC. Conversely, without the ADCs (middle panel) the flux levels have a better match between the arms (spectra were taken at the parallactic angle always). The bottom panel of Fig. 4 shows the XQ-100 spectrum from the middle panel but smoothed and rebinned to SDSS resolution (blue line), and rescaled by a factor of 1.3 to match the corresponding SDSS spectrum (overlaid in red). The good match across wavelengths suggests that slit losses, at least in the SDSS spectral region, are roughly achromatic in the XQ-100 spectra.

Since the accuracy of flux calibrations is unimportant for many of the science applications described in the introduction and an extra exposure might be helpful to increase the S/N, we provide reduced spectra of both observations in these 13 cases and flag them in our database (see Sect. 5).

The remaining observations in the queue proceeded without the ADCs but making sure that the parallactic angle and the lowest possible airmass was chosen.

3. Data reduction

Extraction of NIR spectra can prove a non-trivial task owing to the high sky-background levels. ESO provides a pipeline to reduce X-shooter data, which we have tested. However, in doing so, we noticed that the reduced spectra show systematically large and frequent sky-subtraction residuals in the NIR. Consequently, we opted to implement our own custom pipeline and to reduce XQ-100 data using scripts written in idl by one of us (GDB). Figure 5 shows an example that highlights the differences between the two pipelines in the NIR. Overall, despite some unavoidable residuals, the idl pipeline seems to be more effective than the ESO version available by mid-2014. In the following two sections we describe our pipeline and then provide a qualitative comparison with the ESO version.

thumbnail Fig. 5

Portion of the NIR spectrum of QSO J0003-2603 at z = 4.125 reduced using the X-shooter/Reflex pipeline, version 2.5.2 (top) and our own idl pipeline (bottom). The tickmarks in between spectra indicate Mg ii and Mg i absorption lines in a DLA at z = 3.390. The bottom spectrum is much less affected by the residuals.

3.1. Custom pipeline

The overall reduction strategy is based on the techniques of Kelson (2003), where operations are performed on the un-rectified 2D frames. To achieve this, we generated 2D arrays of slit position and wavelength that served as the coordinate grid for sky modeling and 1D spectrum extraction. A fiducial set of coordinate arrays for each arm was registered to individual science frames using the measured positions of sky and/or arc lines.

Individual frames were bias subtracted (or dark subtracted in the case of the NIR arm) and flat-fielded. The sky emission in each order was then modeled using a b-spline and subtracted. To avoid adding significant extra noise in the NIR arm, composite dark frames were generated from multiple (typically ~ 10) dark exposures with matching integration times. This approach was found to remove the fixed pattern noise in the NIR to the extent that the sky emission could generally be well modeled in each exposure independently, without subtracting a nodded frame, thus avoiding a factor penalty in the background noise. The exception to this was the reddest order (2270−2480 nm), which is problematic because it is vignetted by a baffle designed to mask stray light (see footnote 2 in Sect. 1) This order was therefore nod-subtracted, and the residual sky emission modeled using a b-spline.

Following sky subtraction, the counts in the 2D frames were flux calibrated using response curves generated from observations of spectro-photometric standard stars. Standards observed close in time to the science observations were generally used. For a limited number of objects, however, the temporally closest star was not optimal and unexpected features were observed in the flux-calibrated spectra. In these cases, a fiducial response curve was used to produce an additional flux-calibrated spectrum.

A single 1D spectrum was then extracted simultaneously from across all orders and all exposures of a given object (in a single arm). Extraction was performed on the non-rectified frames to avoid multiple rebinnings and to keep the error correlation across adjacent pixels to a minimum. The number of exposures for each object ranges between 2 and 12, depending on the number of scheduled exposures (two to four) and on the number of times a given OB was executed (typically one, but two or three in cases of interrupted execution and ADC issues). When observations were spread across several nights, a separate 1D spectrum was extracted for each night.

The one-dimensional spectra were binned using a fixed velocity step. This is the only rebinning involved in the reduction procedure. Wavelength bins for the three arms (UVB: 20 km s-1; VIS: 11 km s-1; NIR: 19 km s-1) were chosen to provide roughly 3 pixels per FWHM, taking the nominal X-shooter resolving power for the adopted slits (Table 2). The whole (gap-less) wavelength range is 315 to 1800 nm for spectra taken with the K-band blocking filter, and 315 to 2480 nm for other spectra. Wavelengths were corrected to the vacuum-heliocentric system. When multiple exposures of a single object existed, they were co-added (with the exception of exposures taken with the faulty ADC, which were not included in the co-added spectrum). The stacking was done arm by arm; no attempt was made to merge the arms at this stage, although we do provide joint spectra in the public release (Sect. 5). In the following, we call these reduced data “raw” to distinguish them from the post-processed data (described in Sects. 4 and 5).

Figure 6 shows the distribution of S/N (per pixel) at three different rest-frame wavelengths: 1700 Å (representative of the VIS spectra), 3000 Å (NIR spectra of high-z sources), and 3600 Å (NIR spectra of low-z sources). The respective median signal-to-noise ratios are 33, 25 and 43, as measured in a ± 10 Å window at those rest-frame wavelengths. These values are consistent with the predictions of the X-shooter Exposure Time Calculator, which motivated the setup adopted for the OBs.

Figure B.1 shows all reduced spectra and Fig. 7 shows an example with an expanded wavelength scale.

thumbnail Fig. 6

Distribution of pixel signal-to-noise ratios of the co-added spectra at three different rest-frame wavelengths: 1700, 3000, and 3600 Å. The S/N is computed in a window centered at those wavelengths and spanning ± 10 Å. . The 3 600 Å histogram has fewer elements because not all spectra cover that wavelength (see Sect. 2.2).

3.2. Accuracy of the flux calibration

Comparison with SDSS spectra (expected to have little aperture loss given the 3″ fibers of the spectrograph) shows a systematic underestimation of the flux on the X-shooter part (), mainly due to slit losses induced by the narrow slits used. As shown in the bottom panel of Fig. 4, these slit losses appear to be roughly achromatic.

In some cases the flux values in adjacent arms (especially VIS and NIR) do not match exactly and a gap is observed. In general we expect a mild mismatch which probably depends on seeing, since slit widths are different in each arm and the standard stars, used for flux calibration, are all taken with a 5′′ slit. However, in six spectra a large mismatch is observed (≈ 30% across arms), which cannot be attributed to slit losses only. These spectra are: J0113-2803, J1013+40650, J1524+2123, J1552+1005, J1621−0042, and J1723+2243. For these particular cases, three possible causes were identified: (1) the ADC issue (Sect. 2.2.1); (2) a sudden interruption in the OB execution, which produced UVB and VIS frames with shorter integration time (when this happens, NIR frames are automatically discarded); or (3) problems with flat-fielding. Since an ad hoc treatment of individual targets was beyond the scope of this release and would have compromised the consistency of the reduction process, we decided not to undertake any further action in this direction.

Thus, flux calibration of XQ-100 spectra should not be taken as absolute. The spectral shape is correctly reconstructed and the flux values can be taken as order-of-magnitude estimates, but users of the public data release may want to refer to photometry when an accurate flux measurement is needed.

thumbnail Fig. 7

XQ-100 spectrum of QSO J1117+1311 at z = 3.622, a representative case of the whole sample in terms of S/N. The flux is not corrected for telluric absorption (Sect. 4.1) or rebinned for display purposes. Some emission lines are marked. The red line depicts a manually placed continuum made of cubic splines (see Sect. 4.2 for details). The complete set of XQ-100 spectra is shown in Fig. B.1.

3.3. Comparison with ESO pipeline

The ESO pipeline is run through an environment called Reflex (Freudling et al. 2013), which allows the user to organize the scientific and calibration files and to execute the pipeline in an interactive and graphical fashion. A qualitative comparison between version 2.5.2 of the ESO pipeline and our custom pipeline is as follows:

Wavelength calibration: the ESO pipeline performs a two-stepwavelength calibration of raw spectra, using arc lamp frames. Inthe first step, the positions of the order edges and arc lines arepredicted from a physical model of the instrument. In the secondstep, a 2D mapping from the detector space to the(λ,s) space is computed, where s is the position of the pixel along the slit. This mapping is used to produce the final 2D rectified spectrum. Conversely, our custom-built idl package starts with 2D λ and s coordinate frames that have been carefully calibrated for a single reference exposure, and then shifts these frames to match other exposures using the measured positions of sky (VIS, NIR) or arc (UVB) lines. As a consequence, the cascade for the idl package is simpler and the overall execution time (data retrieval+processing) is generally shorter.

Sky subtraction: both tools implement the Kelson (2003) algorithm for optimal sky subtraction. For reasons that remain unclear, the idl package provides much better results than the ESO pipeline. Residuals of sky-line subtraction in the NIR arm are consistently higher in spectra obtained with the ESO pipeline, as seen in Fig. 5.

Object tracing: in the ESO pipeline, the position of the object is extracted from the 2D rectified (i.e. rebinned) spectrum; in MANUAL mode, the position of the centroid and the trace width are both set constant. The idl package fits the object trace directly on the detector space; when the trace is too faint, it is interpolated from the adjacent orders based on offsets from a standard star trace. Optimal extraction is performed using a variant of the Horne (1986) algorithm.

Coaddition of spectra: the ESO pipeline coadds multiple “nodding” exposures by aligning the object trace in the 2D rectified spectra. Coaddition of already rebinned spectra is not recommended, as it introduces a correlation between the error in adjacent pixels. Conversely, the idl package does not attempt to add the 2D frames. Instead, it optimally extracts a single 1D spectrum from all exposures in the same arm for a given object.

Ease of use: the ESO pipeline can be run automatically through the Reflex interface. The same is true for the idl package, which is easily scriptable. One advantage of the latter is the possibility of obtaining both individual and co-added spectra from an arbitrarily large set of exposures in a single run, shortening the overall execution time.

4. Post-processing

In addition to the (approximately) flux calibrated spectra, we deliver to the community two other higher level science data products: telluric-corrected spectra and QSO continuum fits.

4.1. Removal of telluric features

Telluric absorption affects spectra in both the VIS and NIR arms. Correcting these airmass-dependent spectral features using standard star spectra, even taken relatively close in time to the science targets, can become highly non-trivial owing to the rapidly changing NIR atmospheric transparency. Instead, we opted to derive corrections using model transmission spectra based on the ESO SKYCALC Cerro Paranal Advanced Sky Model (Noll et al. 2012; Jones et al. 2013), version 1.3.5. The SKYCALC models are a function of both airmass and precipitable water vapor (PWV) and span a grid in these parameters providing a spectral resolution of R = 100 000. These corrections were applied to individual-epoch spectra of all XQ-100 sources. Figure 8 shows an example of the results.

thumbnail Fig. 8

Spectrum of QSO J0003−2603 at z = 4.125 before (black) and after (blue) telluric corrections. Three spectral windows with strong telluric absorption are shown. In the middle panel the tickmarks above the spectrum indicate two Fe ii absorption lines, λ2586 and λ2600 Å, associated with the z = 3.390 DLA. In the bottom panel we note how the Mg ii emission line stands out in the corrected spectrum.

Synthetic atmospheric transmission spectra based on the SKYCALC models were fit separately to each VIS and NIR 1D spectrum as a way to remove the observed telluric absorption features. The sky model airmass and PWV parameters, as well as a velocity offset and Gaussian FWHM smoothing kernel, were interactively adjusted for each spectrum in order to minimize the residuals in the model-subtracted spectrum over spectral regions observed to have moderate amounts of absorption, e.g., ~7150−7350 Å, ~7620−7680 Å, ~8120−8350 Å, ~8950−9250 Å, ~9400−9600 Å, ~11 000−11 600 Å, and ~14 600−15 000 Å. Following this initial, interactive parameter selection, an automated parameter selection was performed that searched a grid of only airmass and PWV values in a narrow grid, relative to the best-selected parameters from the interactive search. Multiple sets of best-fit automated parameters were determined for each spectrum by maximizing the S/N measured in the model-subtracted VIS or NIR spectrum over each of the wavelength regions listed above, separately, as well as an average S/N based on all VIS or NIR regions, respectively. The set of parameters used to create the final, telluric-correction model was selected by eye from these multiple, best, model-subtracted spectra.

Owing to the complex nature of correcting for the telluric absorption in this way, which is affected by, for example, the degeneracy between fit parameters and the variable atmospheric conditions during each observation, a single set of parameters generally was not able to optimize the telluric absorption correction at all wavelengths. Similarly, a single quantitative measure of “best” was not attempted. So while the final correction remains somewhat subjective, this allowed an optimization of the correction over the wavelength regions of greatest interest, e.g., near QSO emission lines, such as C iv or Mg ii (see bottom panel of Fig. 8) that will be important for further analysis, and varies for objects at different redshifts.

The telluric correction models were fit to all available 1D spectra. This includes individual-epoch spectra for all objects, as well as spectra co-added from multiple epochs. We note, however, that these model-subtracted, co-added spectra are of poorer quality than those from the individual epochs. This is a result of the coadd of the multiple epochs being done to the pre-telluric-corrected, 2D, unrectified images, which was done to avoid rebinning multiple times. However, this necessarily results in mixed atmospheric features in the co-added spectrum. Such features cannot be cleanly fit by the atmospheric models. In these cases, an argument can be made for coadding the telluric-corrected, 1D spectra instead of the uncorrected 2D frames, even if an additional rebinning is required; however, such post-processing decisions and procedures are left to the user.

4.2. Continuum fitting

For each arm the manually placed continuum was determined by selecting points along the QSO continuum free of absorption (by eye) as knots for a cubic spline. The code used for the continuum fitting is available online7.

For all sightlines, the continuum placement was visually inspected and adjusted such that the final fit resides within the variations of regions with clean continuum. The accuracy of the fits is as good as or better than the S/N of these clean continuum regions. As the continuum fits were created for accurate DLA metal line abundances (Berg et al. 2016), the fits around DLA metal lines have undergone multiple revisions compared to other regions of the spectra. The continuum placement in the Lyα forest is highly subjective due to the lack of clean QSO continuum (e.g. Kirkman et al. 2005), and is particularly difficult to identify around the Lyα absorption of DLAs. The continuum around a DLA Lyα absorption feature in the XQ-100 sightlines requires further refinement on a case-by-case basis to match the N(Hi) fits of the Lyα wings, as implemented in Sánchez-Ramírez et al. (2016).

In regions where the QSO continuum is absorbed, the spline knots were placed at a constant (high) flux at: (i) The Lyman limit if one or more obvious Lyman limits systems are clearly present, and (ii) telluric features (near observed wavelengths 6900 Å, 7600 Å, 9450 Å, 11 400 Å, and 14 000 Å). In some sightlines, there are strong absorption features on top of the Lyα emission line of the QSO, such that the continuum of the emission line is not well constrained (particularly near the peak of the emission). In cases with this strong absorption present, the continuum on the Lyα emission is assumed to follow the interpolation from the cubic spline fit to the surrounding continuum knots.

An example of the continuum presented above is shown in Fig. 7. We note, however, that we provide continua separately for each arm-by-arm spectrum, not for the joint spectra.

5. Description of science data products

All the XQ-100 raw data, along with calibration files are available through the ESO archive8. Advanced science data products (SDP) are publicly available in the form of ESO Phase 3 material9.

The full XQ-100 target list is provided in Table A.1. We also provide a summary file with basic properties (e.g., coordinates and redshifts), spectroscopic properties (e.g., S/N at different rest frame wavelengths), multi-wavelength photometric information, and other spectroscopic data available for each XQ-100 QSO. The detailed content of this summary file is given in Table A.2. We format all the data file names in the same fashion (JNNNNsNNNN) and we provide this standardized name.

Two types of data are provided for each target: (1) individual UVB, VIS, and NIR spectra, also with telluric correction and fitted QSO continuum; and (2) a joint spectrum of the three arms together.

5.1. Individual UVB, VIS, and NIR spectra

There are four different main data files per QSO in the XQ-100 sample: one with the reduced 1D spectrum in the UVB arm, one for the VIS arm, one for the NIR reduced in “stare” mode, and one with the 1D NIR spectrum reduced in nodding mode when available, i.e., when z< 4 (targets observed without the K-band blocking filter10).

Each spectrum file contains wavelength, flux, error on the flux, sky-subtracted flux, and associated error (Sect. 4.1).

When a target was observed more than once (because the observing specifications were not met the first time; see Sect. 2.2), we produced individual spectra of each execution of the OB. Whenever possible, we also produced a co-added spectrum putting together all executions. In the co-added spectra we discarded the first exposures either when they were affected by the ADC issue, or when they were interrupted (as their contribution was negligible due to the short integration time). We define as “primary” spectra those with the best achievable S/N. For targets observed more than once, these correspond to the co-added spectra.

A breakdown of the different spectra provided is shown in Table A.3.

5.2. Joint spectra

Joint spectra contain the three arms merged into a single spectrum. Fluxes from the VIS and NIR arms were rescaled to match the UVB flux level. We first computed the VIS scaling factor (using the UVB-VIS superposition); then, after correcting the VIS, we computed the NIR scaling factor (using the VIS-NIR superposition). In both cases, the scaling factor was defined as the ratio of the two median fluxes in the superposition region. After rescaling, the limit wavelength between UVB and VIS arms was set at 5600 Å and at 10 125 Å between the VIS and NIR arms. The three arms were finally pieced together to create a single spectrum. For targets observed without the K-band blocking filter, the last order of the NIR was taken from the products of the nodding reduction, which are similarly rescaled, cut at 22 700 Å, and pieced together. The resulting spectrum was finally cut in the blue end at 3000 Å and in the red end at 25 000 Å (for targets observed without the K-band blocking filter) and at 18 000 Å (for other targets), to guarantee a comparable wavelength span across the data set. We note that the procedure described above is the result of several choices that may not be appropriate for all scientific analyses.

5.3. Data format

All the spectra we release are binary FITS files. The naming convention is

for the individual arm-by-arm spectra:target_arm_exec.fits;

for the joint spectra: target.fits;

where target is the target name in shortened J2000 coordinates (JNNNN+NNNN or JNNNN−NNNN), arm is the spectral arm, including the optional nodding suffix for the NIR (uvb, vis, nir, or nir_nod), and exec is the optional execution suffix (_1, _2, _3, or blank). The individual arm-by-arm spectrum without the exec suffix is to be regarded as the primary spectrum for the given target in all cases. The list of table columns is

for the individual arm-by-arm spectra: WAVE, FLUX,ERR_FLUX, CONTINUUM, FLUX_TELL_CORR,ERR_FLUX_TELL_CORR;

for the joint spectra: WAVE, FLUX, ERR_FLUX.

The column description is as follows:

WAVE: wavelength in the vacuum-heliocentric system(Å);

FLUX: flux density (erg cm-2 s-1 Å-1);

ERR_FLUX: error of the flux density (erg cm-2 s-1 Å-1);

CONTINUUM: fitted continuum (erg cm-2 s-1 Å-1);

FLUX_TELL_CORR: same as flux, but with the telluric features removed (erg cm-2 s-1 Å-1);

ERR_FLUX_TELL_CORR: error of flux_tc (erg cm-2 s-1 Å-1)

6. Summary

We have presented XQ-100, a legacy survey of 100zem ≃ 3.5−4.5 QSOs observed with VLT/X-shooter. We have provided a basic description of the sample, along with details of the observations, and details of the data reduction process. We have also described the format and organization of the publicly available data, which include spectra corrected for atmospheric absorption and a continuum fit.

XQ-100 provides the first large uniform sample of high-redshift QSOs at intermediate-resolution and with simultaneous rest-frame UV/optical coverage. In terms of number of QSOs this volume represents a 30% increase over the whole extant X-shooter sample. The released spectra are of superb quality, having median S/N ~ 30, 25, and 40 at resolutions of ~ 30−50 km s-1, depending on wavelength. We have indicated that these properties enable a wide range of high-redshift research and soon look forward to seeing the results of this three-year effort in the form of new discoveries and contributions to the field.

1

Available at http://archive.eso.org

6

The number of OB executions is listed in Col. 5 of Table A.2.

10

The overall quality of the nodding reduction is worse than the normal reduction. Their unique advantage is that they extend up to the last NIR order; see Sect. 3 for details.

Acknowledgments

We would like to warmly thank the ESO staff involved in the execution of this Large Programme throughout all its phases. S.L. has been supported by FONDECYT grant number 1140838 and partially by PFB-06 CATA. V.D., I.P., and S.P. acknowledge support from the PRIN INAF 2012 “The X-shooter sample of 100 quasar spectra at z ~ 3.5: Digging into cosmology and galaxy evolution with quasar absorption lines”. SLE acknowledges the receipt of an NSERC Discovery Grant. M.H. acknowledges support by ERC ADVANCED GRANT 320596 “The Emergence of Structure during the epoch of Reionization”. The Dark Cosmology Centre is funded by the Danish National Research Foundation. MVe gratefully acknowledges support from the Danish Council for Independent Research via grant no. DFF – 4002-00275. M.V. is supported by ERC-StG “cosmoIGM”. K.D.D. is supported by an NSF AAPF fellowship awarded under NSF grant AST-1302093. T.S.K. acknowledges funding support from the European Research Council Starting Grant “Cosmology with the IGM” through grant GA-257670. This research has made use of the NASA/IPAC Extragalactic Database (NED), which is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration. Funding for the SDSS and SDSS-II has been provided by the Alfred P. Sloan Foundation, the Participating Institutions, the National Science Foundation, the U.S. Department of Energy, the National Aeronautics and Space Administration, the Japanese Monbukagakusho, the Max Planck Society, and the Higher Education Funding Council for England. The SDSS Web Site is http://www.sdss.org/. The SDSS is managed by the Astrophysical Research Consortium for the Participating Institutions. The Participating Institutions are the American Museum of Natural History, Astrophysical Institute Potsdam, University of Basel, University of Cambridge, Case Western Reserve University, University of Chicago, Drexel University, Fermilab, the Institute for Advanced Study, the Japan Participation Group, Johns Hopkins University, the Joint Institute for Nuclear Astrophysics, the Kavli Institute for Particle Astrophysics and Cosmology, the Korean Scientist Group, the Chinese Academy of Sciences (LAMOST), Los Alamos National Laboratory, the Max-Planck-Institute for Astronomy (MPIA), the Max-Planck-Institute for Astrophysics (MPA), New Mexico State University, Ohio State University, University of Pittsburgh, University of Portsmouth, Princeton University, the United States Naval Observatory, and the University of Washington.

References

  1. Aguirre, A., Schaye, J., Kim, T.-S., et al. 2004, ApJ, 602, 38 [NASA ADS] [CrossRef] [Google Scholar]
  2. Alam, S., Albareti, F. D., Allen de Prieto, C., et al. 2015, ApJS, 219, 12 [NASA ADS] [CrossRef] [Google Scholar]
  3. Becker, R. H., White, R. L., & Helfand, D. J. 1995, ApJ, 450, 559 [NASA ADS] [CrossRef] [Google Scholar]
  4. Becker, G. D., Rauch, M., & Sargent, W. L. W. 2007, ApJ, 662, 72 [NASA ADS] [CrossRef] [Google Scholar]
  5. Becker, G. D., Rauch, M., & Sargent, W. L. W. 2009, ApJ, 698, 1010 [NASA ADS] [CrossRef] [Google Scholar]
  6. Becker, G. D., Sargent, W. L. W., Rauch, M., & Carswell, R. F. 2012, ApJ, 744, 91 [NASA ADS] [CrossRef] [Google Scholar]
  7. Becker, G. D., Bolton, J. S., Madau, P., et al. 2015, MNRAS, 447, 3402 [NASA ADS] [CrossRef] [Google Scholar]
  8. Berg, T., Ellison, S. L., Sánchez-Ramírez, R., et al. 2016, MNRAS, in press [Google Scholar]
  9. Bergeron, J., Petitjean, P., Aracil, B., et al. 2004, The Messenger, 118, 40 [NASA ADS] [Google Scholar]
  10. Brunner, R., Djorgovski, S. G., Prince, T., & Szalay, A. 2002, Massive Data Sets in Astronomy, in Handbook of Massive Data Sets, eds. J. Abello et al. (Dordrecht: Kluwer Academic Publ.), 931 [Google Scholar]
  11. Calverley, A. P., Becker, G. D., Haehnelt, M. G., & Bolton, J. S. 2011, MNRAS, 412, 2543 [NASA ADS] [CrossRef] [Google Scholar]
  12. Capellupo, D. M., Netzer, H., Lira, P., Trakhtenbrot, B., & Mejía-Restrepo, J. 2015, MNRAS. 446. 3427 [Google Scholar]
  13. Chen, H.-W., Wild, V., Tinker, J. L., et al. 2010, ApJ, 724, 176 [NASA ADS] [CrossRef] [Google Scholar]
  14. Croft, R. A. C., Weinberg, D. H., Katz, N., & Hernquist, L. 1998, ApJ, 495, 44 [NASA ADS] [CrossRef] [Google Scholar]
  15. Croft, R. A. C., Weinberg, D. H., Bolte, M., et al. 2002, ApJ, 581, 20 [NASA ADS] [CrossRef] [Google Scholar]
  16. Cutri, R. M., Skrutskie, M. F., van Dyk, S., et al. 2003, 2MASS All Sky Catalog of point sources [Google Scholar]
  17. Dall’Aglio, A., Wisotzki, L., & Worseck, G. 2008, A&A, 491, 465 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  18. De Rosa, G., Venemans, B. P., Decarli, R., et al. 2014, ApJ, 790, 145 [NASA ADS] [CrossRef] [Google Scholar]
  19. Djorgovski, S. G. 2005, Virtual Astronomy, Information Technology, and the New Scientific Methodology, in IEEE Proc. of CAMP05: Computer Architectures for Machine Perception, eds. V. Di Gesu, & D. Tegolo, 125 [Google Scholar]
  20. D’Odorico, V., Cristiani, S., Romano, D., Granato, G. L., & Danese, L. 2004, MNRAS, 351, 976 [NASA ADS] [CrossRef] [Google Scholar]
  21. D’Odorico, V., Bruscoli, M., Saitta, F., et al. 2008, MNRAS, 389, 1727 [NASA ADS] [CrossRef] [Google Scholar]
  22. D’Odorico, V., Calura, F., Cristiani, S., & Viel, M. 2010, MNRAS, 401, 2715 [NASA ADS] [CrossRef] [Google Scholar]
  23. D’Odorico, V., Cupani, G., Cristiani, S., et al. 2013, MNRAS, 435, 1198 [NASA ADS] [CrossRef] [Google Scholar]
  24. Dietrich, M., Appenzeller, I., Vestergaard, M., & Wagner, S. J. 2002, ApJ, 564, 581 [NASA ADS] [CrossRef] [Google Scholar]
  25. Dietrich, M., Hamann, F., Appenzeller, I., & Vestergaard, M. 2003, ApJ, 596, 817 [NASA ADS] [CrossRef] [Google Scholar]
  26. Dietrich, M., Mathur, S., Grupe, D., & Komossa, S. 2009, ApJ, 696, 1998 [NASA ADS] [CrossRef] [Google Scholar]
  27. Eisenstein, D. J., Weinberg, D. H., Agol, E., et al. 2011, AJ, 142, 72 [Google Scholar]
  28. Flesch, E. 2015, PASA, 32, 10 [Google Scholar]
  29. Freudling, W., Romaniello, M., Bramich, D. M., et al. 2013, A&A, 559, A96 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  30. Hamann, F., & Ferland, G. 1999, ARA&A, 37, 487 [NASA ADS] [CrossRef] [Google Scholar]
  31. Hamann, F., Korista, K. T., Ferland, G. J., Warner, C., & Baldwin, J. 2002, ApJ, 564, 592 [NASA ADS] [CrossRef] [Google Scholar]
  32. Ho, L. C., Goldoni, P., Dong, X., Greene, J. E., & Ponti, G. 2012, ApJ, 754, 11 [NASA ADS] [CrossRef] [Google Scholar]
  33. Horne, K. 1986, PASP, 98, 609 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  34. Iršič, V., Viel, M., Berg, T. A. M., et al. 2016, MNRAS, submitted [Google Scholar]
  35. Irwin, M., McMahon, R. G., & Hazard, C. 1991, ASP Conf. Ser., 21, 117 [NASA ADS] [Google Scholar]
  36. Jiang, L., Fan, X., Vestergaard, M., et al. 2007, AJ, 134, 1150 [NASA ADS] [CrossRef] [Google Scholar]
  37. Jones, A., Noll, S., Kausch, W., Szyszka, C., & Kimeswenger, S. 2013, A&A, 560, 91 [Google Scholar]
  38. Kelson, D. D. 2003, PASP, 115, 688 [NASA ADS] [CrossRef] [Google Scholar]
  39. Kim, T.-S., Carswell, R. F., Cristiani, S., D’Odorico, S., & Giallongo, E. 2002, MNRAS, 335, 555 [NASA ADS] [CrossRef] [Google Scholar]
  40. Kirkman, D., Tytler, D., Suzuki, N., O’Meara, J. M., & Lubin, D. 2003, ApJS, 149, 1 [NASA ADS] [CrossRef] [Google Scholar]
  41. Kirkman, D., Tytler, D., Suzuki, N., et al. 2005, MNRAS, 360, 373 [Google Scholar]
  42. Kondo, S., Kobayashi, N., Minami, A., et al. 2008, in Panoramic Views of Galaxy Formation and Evolution, eds. T. Kodama, T. Yamada, & K. Aoki, ASP Conf. Ser., 399, 209 [Google Scholar]
  43. Ledoux, C., Petitjean, P., & Srianand, R. 2003, MNRAS, 346, 209 [NASA ADS] [CrossRef] [Google Scholar]
  44. Lu, L., Sargent, W. L. W., Barlow, T. A., Churchill, C. W., & Vogt, S. S. 1996, ApJS, 107, 475 [NASA ADS] [CrossRef] [Google Scholar]
  45. Marziani, P., Sulentic, J. W., Stirpe, G. M., Zamfir, S., & Calvani, M. 2009, A&A, 495, 83 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  46. Matejek, M. S., & Simcoe, R. A. 2012, ApJ, 761, 112 [NASA ADS] [CrossRef] [Google Scholar]
  47. Ménard, B., Wild, V., Nestor, D., et al. 2011, MNRAS, 417, 801 [NASA ADS] [CrossRef] [Google Scholar]
  48. Molaro, P., Centurión, M., Whitmore, J. B., et al. 2013, A&A, 555, A68 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  49. Murphy, M. T., Webb, J. K., & Flambaum, V. V. 2003, MNRAS, 345, 609 [NASA ADS] [CrossRef] [Google Scholar]
  50. Noll, S., Kausch, W., Barden, M., et al. 2012, A&A, 543, A92 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  51. Noterdaeme, P., Laursen, P., Petitjean, P., et al. 2012a, A&A, 540, A63 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  52. Noterdaeme, P., Petitjean, P., Carithers, W. C., et al. 2012b, A&A, 547, L1 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  53. O’Meara, J. M., Lehner, N., Howk, J. C., et al. 2015, AJ, 150, 111 [NASA ADS] [CrossRef] [Google Scholar]
  54. Palanque-Delabrouille, N., Yèche, C., Borde, A., et al. 2013, A&A, 559, A85 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  55. Pâris, I., Petitjean, P., Aubourg, É., et al. 2012, A&A, 548, A66 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  56. Pâris, I., Petitjean, P., Aubourg, É., et al. 2014, A&A, 563, A54 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  57. Patat, F., & Hussain, G. A. J. 2013, in Organizations, People and Strategies in Astronomy 2 (OPSA 2, Venngeist), ed. A., Heck, 231 [Google Scholar]
  58. Péroux, C., Bouché, N., Kulkarni, V. P., York, D. G., & Vladilo, G. 2011 MNRAS, 410, 2237 [Google Scholar]
  59. Perrotta, S., D’Odorico, V., Prochaska, J. X., et al. 2016, MNRAS, in press [Google Scholar]
  60. Prochaska, J. X., Gawiser, E., Wolfe, A. M., Castro, S., & Djorgovski, S. G. 2003, ApJ, 595, 9 [Google Scholar]
  61. Prochaska, J. X., Worseck, G., & O’Meara, J. M. 2009a, ApJ, 705, 113 [Google Scholar]
  62. Prochaska, J., Xavier, W., & Arthur, M. 2009b, ApJ, 696, 1543 [NASA ADS] [CrossRef] [Google Scholar]
  63. Prochaska, J. X., O’Meara, J. M., & Worseck, G. 2010, ApJ, 718, 392 [NASA ADS] [CrossRef] [Google Scholar]
  64. Rafelski, M., Wolfe, A. M., Prochaska, J. X., Neeleman, M., & Mendez, A. J. 2012, ApJ, 755, 89 [NASA ADS] [CrossRef] [Google Scholar]
  65. Rudie, G. C., Steidel, C. C., Trainor, R. F., et al. 2012, ApJ, 750, 67 [NASA ADS] [CrossRef] [Google Scholar]
  66. Ryan-Weber, E. V., Pettini, M., Madau, P., & Zych, B. J. 2009, MNRAS, 395, 1476 [NASA ADS] [CrossRef] [Google Scholar]
  67. Sánchez-Ramírez, R., Ellison, S. L., Prochaska, J. X., et al. 2016, MNRAS, 456, 4488 [NASA ADS] [CrossRef] [Google Scholar]
  68. Scannapieco, E., Pichon, C., Aracil, B., et al. 2006, MNRAS, 365, 615 [NASA ADS] [CrossRef] [Google Scholar]
  69. Schaye, J., Theuns, T., Rauch, M., Efstathiou, G., & Sargent, W. L. W. 2000, MNRAS, 318, 817 [NASA ADS] [CrossRef] [Google Scholar]
  70. Schneider, D. P., Richards, G. T., Hall, P. B., et al. 2010, AJ, 139, 2360 [NASA ADS] [CrossRef] [Google Scholar]
  71. Simcoe, R. A., Cooksey, K. L., Matejek, M., et al. 2011, ApJ, 743, 21 [NASA ADS] [CrossRef] [Google Scholar]
  72. Songaila, A. 2005, AJ, 130, 1996 [NASA ADS] [CrossRef] [Google Scholar]
  73. Songaila, A., & Cowie, L. L. 2010, ApJ, 721, 1448 [NASA ADS] [CrossRef] [Google Scholar]
  74. Srianand, R., Chand, H., Petitjean, P., & Aracil, B. 2004, Phys. Rev. Lett., 92, 121302 [NASA ADS] [CrossRef] [PubMed] [Google Scholar]
  75. Storrie-Lombardi, L. J., McMahon, R. G., Irwin, M. J., & Hazard, C. 1994, ApJ, 427, 13 [Google Scholar]
  76. Sulentic, J. W., Stirpe, G. M., Marziani, P., et al. 2004, A&A, 423, 121 (Paper I) [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  77. Sulentic, J. W., Repetto, P., Stirpe, G. M., et al. 2006, A&A, 456, 929 (Paper II) [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  78. Suzuki, N., Tytler, D., Kirkman, D., O’Meara, J. M., & Lubin, D. 2005, ApJ, 618, 592 [NASA ADS] [CrossRef] [Google Scholar]
  79. Vernet, J., Dekker, H., D’Odorico, S., et al. 2011, A&A, 536, A105 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  80. Vestergaard, M., & Osmer, P. S. 2009, ApJ, 699, 800 [Google Scholar]
  81. Vestergaard, M., & Peterson, B. M. 2006, ApJ, 641, 689 [NASA ADS] [CrossRef] [Google Scholar]
  82. Viel, M., Haehnelt, M. G., & Springel, V. 2004, MNRAS, 354, 684 [NASA ADS] [CrossRef] [Google Scholar]
  83. Viel, M., Bolton, J. S., & Haehnelt, M. G. 2009, MNRAS, 399, L39 [NASA ADS] [CrossRef] [Google Scholar]
  84. Viel, M., Becker, G. D., Bolton, J. S., & Haehnelt, M. G. 2013, Phy. Rev. D, 88, 043502 [NASA ADS] [CrossRef] [Google Scholar]
  85. Worseck, G., & Prochaska, J. X. 2011, ApJ, 728, 23 [NASA ADS] [CrossRef] [Google Scholar]
  86. Worseck, G., Prochaska, J. X., O’Meara, J. M., et al. 2014, MNRAS, 445, 1745 [NASA ADS] [CrossRef] [Google Scholar]
  87. Wright, E. L., Eisenhardt, P. R. M., Mainzer, A. K., et al. 2010, AJ, 140, 1868 [NASA ADS] [CrossRef] [Google Scholar]
  88. Wolfe, A. M., Gawiser, E., & Prochaska, J. X. 2005, ARA&A, 43, 861 [NASA ADS] [CrossRef] [MathSciNet] [Google Scholar]
  89. York, D. G., Adelman, J., Anderson, Jr., J. E., et al. 2000, AJ, 120, 1579 [Google Scholar]
  90. Zafar, T., Popping, A., & Péroux, C. 2013, A&A, 556, A140 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  91. Zhu, G., & Ménard, B. 2013, ApJ, 770, 130 [NASA ADS] [CrossRef] [Google Scholar]
  92. Zuo, W., Wu, X-B., Fan, X., et al. 2015, ApJ, 799, 189 [NASA ADS] [CrossRef] [Google Scholar]

Appendix A: Additional tables

Table A.1

Summary of XQ-100 target properties.

Table A.2

Parameters associated with each XQ-100 object in the public repository.

Table A.3

Number of reduced spectra.

Appendix B: Spectra

thumbnail Fig. B.1

XQ-100 spectra. Names follow the XQ-100 convention (Sect. 5); see Table A.1 for a correspondence with literature names. Emission redshifts were estimated using the result of a principal component analysis (Pâris et al. 2012). The flux has been smoothed with a five-pixel median filter for displaying purposes.

thumbnail Fig. B.1

continued.

thumbnail Fig. B.1

continued.

thumbnail Fig. B.1

continued.

thumbnail Fig. B.1

continued.

thumbnail Fig. B.1

continued.

thumbnail Fig. B.1

continued.

thumbnail Fig. B.1

continued.

thumbnail Fig. B.1

continued.

thumbnail Fig. B.1

continued.

All Tables

Table 1

Requested observing conditions.

In the text
Table 2

Instrument setup.

In the text
Table A.1

Summary of XQ-100 target properties.

In the text
Table A.2

Parameters associated with each XQ-100 object in the public repository.

In the text
Table A.3

Number of reduced spectra.

In the text

All Figures

thumbnail Fig. 1

Sky distribution of XQ-100 sources. The color scale indicates emission redshifts.

In the text
thumbnail Fig. 2

XQ-100 emission redshifts.

In the text
thumbnail Fig. 3

XQ-100 R-magnitudes (APM).

In the text
thumbnail Fig. 4

XQ-100 spectra of the same QSO, J1126−0124, taken with the faulty ADCs in April 2012 (top panel) and repeated with the disabled ADCs in February 2014 (middle panel). Both observations were executed at a similar airmass of ≈ 1.3 at the parallactic angle. The dashed lines indicate the boundaries of the X-shooter arms. The match is better between the VIS and NIR arms in the middle panel. The bottom panel shows the same February 2014 XQ-100 spectrum but smoothed and rebinned to SDSS resolution (blue line), and rescaled by a factor of 1.3 to match the corresponding SDSS spectrum (overlaid in red). The good match suggests that slit losses in the XQ-100 data are roughly achromatic.

In the text
thumbnail Fig. 5

Portion of the NIR spectrum of QSO J0003-2603 at z = 4.125 reduced using the X-shooter/Reflex pipeline, version 2.5.2 (top) and our own idl pipeline (bottom). The tickmarks in between spectra indicate Mg ii and Mg i absorption lines in a DLA at z = 3.390. The bottom spectrum is much less affected by the residuals.

In the text
thumbnail Fig. 6

Distribution of pixel signal-to-noise ratios of the co-added spectra at three different rest-frame wavelengths: 1700, 3000, and 3600 Å. The S/N is computed in a window centered at those wavelengths and spanning ± 10 Å. . The 3 600 Å histogram has fewer elements because not all spectra cover that wavelength (see Sect. 2.2).

In the text
thumbnail Fig. 7

XQ-100 spectrum of QSO J1117+1311 at z = 3.622, a representative case of the whole sample in terms of S/N. The flux is not corrected for telluric absorption (Sect. 4.1) or rebinned for display purposes. Some emission lines are marked. The red line depicts a manually placed continuum made of cubic splines (see Sect. 4.2 for details). The complete set of XQ-100 spectra is shown in Fig. B.1.

In the text
thumbnail Fig. 8

Spectrum of QSO J0003−2603 at z = 4.125 before (black) and after (blue) telluric corrections. Three spectral windows with strong telluric absorption are shown. In the middle panel the tickmarks above the spectrum indicate two Fe ii absorption lines, λ2586 and λ2600 Å, associated with the z = 3.390 DLA. In the bottom panel we note how the Mg ii emission line stands out in the corrected spectrum.

In the text
thumbnail Fig. B.1

XQ-100 spectra. Names follow the XQ-100 convention (Sect. 5); see Table A.1 for a correspondence with literature names. Emission redshifts were estimated using the result of a principal component analysis (Pâris et al. 2012). The flux has been smoothed with a five-pixel median filter for displaying purposes.

In the text
thumbnail Fig. B.1

continued.
In the text
thumbnail Fig. B.1

continued.
In the text
thumbnail Fig. B.1

continued.
In the text
thumbnail Fig. B.1

continued.
In the text
thumbnail Fig. B.1

continued.
In the text
thumbnail Fig. B.1

continued.
In the text
thumbnail Fig. B.1

continued.
In the text
thumbnail Fig. B.1

continued.
In the text
thumbnail Fig. B.1

continued.
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.