Extreme UV QSOs
We present a sample of spectroscopically confirmed QSOs with FUV-NUV color (as measured by GALEX photometry, FUV band: 1344-1786Å, NUV band: 1771 - 2831Å) bluer than canonical QSO templates and than the majority of known QSOs. We analyze their FUV to NIR colors, luminosities and optical spectra. The sample includes a group of 150 objects at low redshift (z 0.5), and a group of 21 objects with redshift 1.7z2.6. For the low redshift objects, the “blue” FUV-NUV color may be caused by enhanced Ly emission, since Ly transits the GALEX FUV band from z=0.1 to z=0.47. Synthetic QSO templates constructed with Ly up to 3 times stronger than in standard templates match the observed UV colors of our low redshift sample. Optical photometric and spectroscopic properties of these QSOs are not atypical. The H emission increases, and the optical spectra become bluer, with increasing absolute UV luminosity. The lack of selected objects at intermediate redshift is consistent with the fact that for z=0.481.63, Ly is included in the GALEX NUV band, making the observed FUV-NUV redder than the limit of our sample selection. The UV-blue QSOs at redshift 2, where the GALEX bands sample restframe 450-590Å (FUV) and 590-940Å (NUV), are fainter than the average of UV-normal QSOs at similar redshift in NUV, while they have comparable luminosities in other bands. Therefore we speculate that their observed FUV-NUV color may be explained by a combination of steep flux rise towards short wavelengths and dust absorption below the Lyman limit, such as from small grains or crystalline carbon (nanodiamonds). The ratio of Ly to CIV could be measured in 10 objects; it is higher (30% on average) than for UV-normal QSOs, and close to the value expected for shock or collisional ionization. However, optical spectra are taken at different times than the UV photometry, which may bias the comparison if lines are variable. These QSO groups are uniquely set apart by the GALEX photometry within larger samples, given that their optical properties are not unusual.
In all QSO samples there is concern that selection effects are present and significant, particularly in whether whole classes of objects are not included, or even known. This study aims at characterizing a population of objects with rising fluxes at UV observed wavelengths. Following our work of classification of UV sources from the GALEX 111 The Galaxy Evolution Explorer, GALEX (Martin et al. 2005), is a NASA Small Explorer performing imaging surveys of the sky in two UV bands simultaneously: FUV (1344 - 1786 Å, = 1528 Å) and NUV (1771 - 2831 Å, = 2271 Å) with different coverage and depth. See Bianchi (2008) for a summary of the UV sources classification and statistics in the main surveys and Morrissey et al. (2007) for instrument description and performance. sky surveys (Bianchi 2008, Bianchi et al. 2006, 2007, 2008 and references therein), we have suspected the existence of a substantial number of extragalactic objects with FUV-NUV color much bluer (more negative) than canonical QSO templates and than the majority of QSOs in known samples. Such objects are rather “normal” at optical wavelengths (spectroscopically and photometrically) but they stand out in the observed UV range, having FUV-NUV colors similar to those of hot white dwarfs (WD). Photometrically, these objects have UV-to-optical colors similar to a stellar binary containing a hot WD and a cooler companion. That a significant number of “FUV-NUV”-blue extragalactic objects existed was first suspected by Bianchi et al. (2007), based on density counts of photometrically selected WD candidates. In fact, the number of objects per square degree whose SED (FUV to near-IR) is consistent with a single hot WD increases with magnitude down to m21 (AB) and then declines, consistent with Milky Way models. However, the density of objects with similarly blue UV color but redder optical colors, that we would expect to be hot WDs with a cool companion, increases considerably at fainter magnitudes, suggesting that a significant number of faint extragalactic objects may be included in the color-color locus of these stellar binaries (Bianchi et al. 2007, Bianchi 2008). In this work we focus on these QSOs, which display very blue observed FUV-NUV colors, and investigate whether their properties are unlike those of known objects.
2 Sample and data.
The sample was extracted from the catalog of matched UV/optical sources of Bianchi (2008), obtained by matching the UV sources in the GALEX third data release (GR3)222GR3 is available from the MAST archive at http://galex.stsci.edu , to the Sloan Digital Sky Survey (SDSS) sixth data release (DR6). GALEX provides sky surveys with different sky area coverage and depth: we restricted this work to the “Medium Imaging Survey” (MIS) data, which reaches a typical ABmag of 22.7 in both FUV and NUV. The overlap area between GALEX-GR3 MIS data and SDSS-DR6 is 573 square degrees (Bianchi 2008), taking into account that only the central 1 degree diameter part of the GALEX fields was used in our master catalog, to assure homogeneous photometry quality and exclude defects in the outer parts of the circular field. For each matched source, GALEX provides FUV (1344 - 1786 Å, = 1528 Å) and NUV (1771 - 2831 Å, = 2271 Å) photometry, and the SDSS provides u, g, r, i, z photometry. More details on the matchings procedure and the catalog are given by Bianchi (2008), Bianchi et al (2009).
In order to characterize the suspected “FUV-NUV blue” QSOs, we extracted from the matched UV/optical source catalog of Bianchi (2008) the spectroscopically confirmed QSOs with FUV-NUV 0.1 (AB mag): this FUV-NUV limit is “bluer”(more negative) than the synthetic FUV-NUV color from the two QSO canonical templates used by Bianchi et al. (2007), which represent average QSO properties, at any redshift. The colors of the canonical templates are shown in Fig. 1 and 2 (cyan diamonds) as a function of redshift. We will refer to this sample as “UV-blue” QSOs for brevity throghout the paper. It is restricted to sources with photometric errors smaller than 0.3 mag in both FUV and NUV, and color FUV-NUV 0.1, for which SDSS spectroscopy exists and gives a “QSO” classification. The requirement of available SDSS spectroscopy effectively limits the sample to brighter magnitudes, but it provides a classification and useful information, which will help interpreting larger samples of photometric candidates. Such relatively bright objects will also be accessible to the spectroscopic capabilities of the refurbished HST, and to other follow-up observations. Note that the SDSS spectroscopic class “QSO” (class 4) probably includes also Seyfert galaxies. We will use here the generic term “QSO” to reflect our selection criterion from the SDSS spectroscopic database. It is important to note that spectroscopic targets in the SDSS were selected with criteria unrelated to our present objective and therefore our UV-blue spectroscopically confirmed QSOs may be a biased sub-sample among the UV-blue QSO photometric candidates.
These selection criteria produced an initial sample of 174 objects. One additional object was excluded because its u-band measurement is saturated. The photometric properties of the sample QSOs are presented in section 3, and their optical spectra and overall properties are analyzed in section 4. The selected objects are shown in two color-color diagrams, Figs. 1 and 2, where they can be compared with other classes of objects, in particular hot stars, typical QSOs, and galaxies. Our analysis of the spectra (section 4) generally confirms the redshift measurement from the pipeline. However, we found that one object (GALEX J172101.08+532433.7, RA=260.2544916, Dec=53.4093516, SDSS match id= 587725490527731868) was misclassified by the SDSS spectroscopic pipeline as a QSO with redshift z=2.7: its spectrum is that of a hot star. It is shown in some figures because it is interesting to note its position in the color-color diagrams (Figs 1 and 2): the GALEX photometry clearly place this object on the stellar sequence and not as a QSO candidate. Coordinates, photometry, and other relevant information are given in Table 1, and sample images are shown in Fig. 3.
The SDSS optical spectra (range 3800-9200Å, resolution 1800), provide the initial classification as QSOs and a measure of redshift for the objects. Our sample includes a group of 151 objects at low redshift (0.041z 0.436), and 21 objects with redshift between 1.7 and 2.6, all pointlike. Only one object has intermediate redshift (z=0.93) and its identification as a QSO is dubious. It has fairly large photometric errors: FUV=21.490.12, NUV=21.430.10, but the typical QSO FUV-NUV color at this redshift is much redder (by 1 mag, see Fig. 1); therefore if it is a QSO it would be quite anomalous. At this redshift the GALEX bands sample rest wavelengths of 1300Å(NUV) and 800Å(FUV), and a very blue FUV-NUV color would be not expected. The image and spectrum of this object are shown in Fig. 4 -bottom. The spectrum shows one emission line that is identified as MgII for the alleged redshift and possibly a few absorptions including one at the red end that could be H. The observed wavelengths of the lines are not obvious for identification, assuming other values of redshift the observed line(s) might be CIII] 1909 or [OII]3727, but then other lines such as CIV 1550 or [OIII] 5007 should be present and are not. So, either z=0.93 is right or perhaps the emission is some artifact in the spectrum. Therefore, we consider this object doubtful and do not include it in our analysis. The lack of objects between redshift 0.5 and 1.7 is consistent with our selection of very blue FUV-NUV color, because Ly is in the NUV band between z=0.48 and z=1.63, causing brighter flux in NUV and consequently much redder FUV-NUV color, as can be seen in Fig. 1.
We also caution that while the GALEX FUV and NUV images are taken simultaneously, the SDSS imaging was taken at a different time from the GALEX observations, therefore any significant variability may affect the combined UV and optical colors, such as NUV-r. For this reason we based our initial sample selection on the FUV-NUV color only. Some of our targets have repeated observations with GALEX, but most repeated measurements are from the AIS (All-sky Imaging Survey), which has about 10 times shorter exposures than MIS (used in this work) and therefore large photometric uncertainties. In a few cases repeated measurements are discrepant by 2 in the combined photometric errors: however most of the discrepant measurements have artifact flags set. We compile for completeness all repeated measurements with exposures longer than 400 sec and formally discrepant by 2 in Table 2, where we provide also comments that help assess reliability, based on the flags from the pipeline photometry, and our visual inspection of the images. We only excluded measurements on the very edge of the GALEX field (flag “rim”) however we did not apply error cuts nor area cut, for the purpose of an exhaustive comparison, while our analysis sample is restricted to measurements in the central 0.5 deg. radius of the field for accurate photometry (Sect. 2). In a few cases the discrepancies in the repeated measurements cannot be ascribed to artifacts, and these objects may deserve dedicated follow-up photometry. In some cases the variation affects the FUV-NUV color, and in particular some repeated measurements have redder FUV-NUV than our initially selected dataset. All discrepant repeated measurements with MIS exposures (2 high-redshift objects and 10 low-redshift objects) have FUV-NUV0.1. If we consider also AIS data (exposure times 100 sec), we find 55 additional repeated measurements discrepant by 2 , of which 36 give FUV-NUV redder than our selected dataset (MIS measurements), and other 19 bluer. Fast variability is not unknown in QSOs, and in particular line strength may vary on short time scales, while it would be less plausible for dust effects to change rapidly. We stress that a variabilty assessment however would require custom photometry, and while the standard pipeline photometry is good for statistical analysis, such as the scope of this work, we should refrain from overinterpreting individual measurements and individual variations. Other 72 AIS and 7 MIS repeated measurements agree within 2 with our selected measurements given in Table 1. One object in the initial sample, although part of the MIS survey, has a 40 sec exposure in FUV and 1518 sec in NUV: although its FUV error (0.23) meets our selection limits, a longer MIS exposure of 853 sec in both FUV and NUV and smaller errors gives FUV-NUV=0.17, so it is eliminated from our analysis sample.
A larger sample of about 30,000 QSOs candidates with “normal” FUV-NUV colors (i.e. similar to the standard template), will be presented elsewhere. We will refer to this sample as “UV-normal” in the discussion of the UV-blue sample for comparison purposes.
3 General photometric properties
Of the 174 sample objects, 64 sources are classified as point-like (at the resolution of the SDSS imaging, 1.4) and 110 as extended (all at low redshift), by the SDSS pipeline. We will keep the pipeline classification because it is derived from an objective procedure, although the result depends on the contrast between central source and underlying galaxy. GALEX and SDSS imaging for a subsample of objects, presented in Fig. 3, shows that the definition of “pointlike” (P) or “extended” (E) is not clear-cut.
The sample selection, as described in section 1, was restricted to MIS sources with photometric errors less than 0.3 mag in FUV and NUV. Of the 64 point-like sources, 42/33 have errors less than 0.1 mag/0.05 mag, and only 4 have errors between 0.2 and 0.3 mag, in FUV. As for the NUV measurements, 36/45 pointlike sources have errors 0.05/0.1 mag, and 10 have errors larger than 0.2 mag. In the r-band, all but 2 objects have errors smaller than 0.05 mag (one object has an error of 0.14 mag, and one of 0.08 mag). Of the 110 extended objects, 108 have r-band error 0.04 mag, 97/94 have NUV / FUV error 0.1 mag. Sources with large photometric errors are identified in the figures. Most objects with the larger errors are in the z2 group. At this redshift, our color cut of FUV-NUV0.1 is more than half a magnitude bluer than the average value for UV-normal QSOs (e.g. Fig.1), therefore these objects may be truly extreme with respect to average samples in spite of their large photometric errors. We point out that the object GALEX J113223.4+641958 (SDSS J113223.42+641958.4) has a u-band magnitude of 28.70.46 (petromag) while the magnitude listed on the explorer page for this object is u=25.073.05. It has no artifact flags, the pipeline records only a warning “no petrosian radius could be determined. Petrosian magnitude still usable; the object is blended with an extended object”. The surrounding galaxy can be seen in the SDSS imaging with a radius of about 5. We regard the petrosian magnitude as unreliable in the u-band. Magnitudes in other bands for this object have smaller errors and seem more consistent among measurements. All other objects have u-band magnitudes brighter than 22.2, consistent with the SDSS limit (see Fig. 3 of Bianchi et al. 2007). We give in Table 1 petrosian magnitude measurements for the SDSS data, for consistency among the sample and with other extragalactic works. The SDSS pipeline also provides magnitudes measured in different ways: psf fitting, DeVaucouleurs model, and exponential fitting. A description of the different magnitudes can be found on the SDSSS web site http://www.sdss.org/dr5/algorithms/photometry.html. We checked for all objects whether the different measurements are discrepant. As expected, for pointlike sources the average difference is within the 1 errors and the largest discrepancies close to 3. Disagreement between psf-mag and petromag tend to increase at longer wavelengths, where the extended galaxy is contributing. For extended objects, the measurements from petrosian and deVaucouleur-profile fitting agree on average within better than 2, while psf magnitudes are more discrepant as expected and should not be used.
The FUV-NUV and NUV-g colors of the sample objects are plotted as a function of redshift in Fig. 5, and the FUV, NUV, and r-band magnitudes in Fig. 6. “Extended” sources are plotted with different symbols, to explore possible trends, although the classification must be regarded only as an indication as pointed out above. Photometric errors (1 error bars are shown) in most cases are quite small compared with the spread in FUV-NUV color observed in our sample. Fig. 5 shows that the high-redshift objects have a wider range of FUV-NUV color than the low-redshift point-like sample, although the spread may simply be caused by the large errors of these faint objects. The lack of extended objects at high redshift is likely due to the fact that for these more distant objects the same imaging does not reveal the underlying galaxy. This question will be investigated with deeper imaging aimed at revealing the underlying galaxy in the distant objects and to probe the contrast to the central source (Hutchings, Scholz, and Bianchi 2009).
Fig. 6 shows that low-redshift pointlike QSOs tend to be brighter than extended ones, in both FUV and NUV. In the r-band, however, the magnitude spread is less (about 4 mags across the sample) and no preferential distribution is seen between pointlike and extended samples. Low redshift pointlike objects are also brighter (observed magnitudes) than higher redshift objects by about 2-3 magnitudes, but their intrinsic luminosity is lower. The distribution of observed magnitudes (left panels) is useful for comparison with other samples, and to estimate the possible contamination by these objects to density counts of other UV-blue objects such as Milky Way WDs, which have similar FUV-NUV colors (see Bianchi et al. 2007, 2008, and Fig. 1), as well as for planning follow-up observations. The misclassified star is shown in these panels. In the right-side panels of Fig. 6 the absolute magnitudes are plotted (the luminosity distance was derived from the redshift using standard cosmology H = 70 km/s/Mpc, =0.3, = 0.7 ); the distant objects are intrinsically more luminous. We plotted for comparison the median absolute magnitudes of our UV-normal QSO comparison sample (solid line in the right side plots). The high redshift UV-blue QSOs have luminosities similar to UV-normal QSOs, except in NUV, where they are fainter. The comparison suggests some absorption in the NUV band (restframe 900Å for z=2) as an explanation of their FUV-NUV color. We will discuss this point later.
4 Analysis and Discussion.
We discuss the two groups, z0.5, and z2 QSOs, separately because the FUV and NUV bands sample different restframe spectral regions and therefore the explanations for their blue FUV-NUV colors are different.
4.1 The low redshift QSOs
The majority of our analysis sample has redshift 0.5 (150 objects, 109 extended and 41 pointlike). Ly transits the GALEX FUV band from z=0.1 to z=0.47, and this fact suggests that an intense Ly emission may be the cause for the “FUV excess” of these objects. To test this hypothesis, we constructed templates with Ly emission enhanced relative to standard templates, and derived their synthetic broad-band colors. Such ad hoc templates with Ly enhanced by up to 3 match the range of observed FUV-NUV colors of our low-z sample, and are shown in Figs. 1 and 2 (dark blue diamonds) together with synthetic colors from canonical templates (cyan diamonds), as well as in Fig. 5(top). Note from these figures that our simple cut of FUV-NUV0.1 produced a low-redshift sample bluer in FUV-NUV than standard templates with a spread of about half a magnitude: the redder objects among our sample are very close to the standard template at z0.2, the UV-bluest objects differ by up to .5 mag (one by 1 mag) and are concentrated around z0.2 where Ly is at the peak of the FUV filter transmission. The modulation with redshift of the hypothetical enhanced-Ly effect, due to the filter transmission, is seen in the ad hoc template plotted in Fig. 5.
None of our sample objects have UV spectra, which would directly reveal the cause of their blue FUV-NUV color. We examined their optical spectra and in particular H, the strongest line in all the objects. Fig. 7 shows the optical spectra of our low redshift sample, stacked, and compared with the standard template (cyan). The majority of the pointlike sources have emission lines stronger than the average template, and bluer spectral slope (flux increasing at shorter wavelengths). For the extended sources, however, line strength is generally typical and the spectral slope mostly redder than the standard template, reflecting the non-negligible contribution by the underlying galaxy (SDSS spectra are taken through a 3 diameter aperture). Sample spectra for both pointlike and extended QSOs are also shown in Fig. 8. There is a wide range of line strengths and profiles, as well as spectral slopes.
The SDSS pipeline provides automated measurements of width and equivalent width (EW) of the major lines, performed with line fitting; we downloaded and examined those quantities. We found that the centering of the line could be used, while line width and equivalent width from the pipeline are not reliable for most spectra (examples in Fig. 9). We remeasured the H line, first by hand to assess the difference from the pipeline measurements, and then with an ad hoc algorithm for more objective results. The line width estimated by our code is also shown in Fig. 9. In order to minimize the complication of narrow absorptions and emissions in some profiles, we did not measure the width at half maximum (peak) but the width at the average flux value of the total line emission. We consider our measurements more homogeneous than the pipeline values, as shown in Fig. 9, and we use them in the following analysis. H width, EW and fluxes (F) measured (at restframe wavelengths) for the low-z sample are reported in Table 3. Errors from the spectra S/N and continuum placement uncertainties, are estimated to be less than 10%. As a further check, measurements from our code agree with our manual measurements with by-eye location of the continuum to a few percent in all but a few cases, where they agree within 10%. Our measurements include the entirety of the emission feature, no attempt was made to separate narrow components when present, and no correction for [NII] was applied.
We searched for possible correlations of H intensity and width, and of the optical spectral slope, with the UV color and absolute luminosity. The spectral slope was measured as the ratio of fluxes integrated in two intervals which are rather free of conspicuous features in most spectra: rest wavelengths 3500-3700Å and 6000-6400Å. We compared several quantities, and we show six interesting cases in Fig. 10. No obvious correlation is seen with the FUV-NUV color. The H flux, and EW, increase with absolute UV luminosity, and to a much lesser extent with u-band luminosity, but not with luminosity at longer wavelengths. Similarly, the optical spectral slope possibly correlates with H EW and with UV luminosity (but not with optical luminosity: the r-band is also shown in Fig.10): it becomes bluer for brighter UV luminosities, suggesting that for low QSO luminosity the galaxy relative contribution is more significant. There is a clear difference between pointlike and extended sources: the latter have a flatter (redder) optical slope and lower H emission, reflecting the contribution of the host galaxy. This result emphasizes the role of UV studies in extending the known properties of QSOs. If we restrict the sample around redshift z=0.2, where we have a wider FUV-NUV observed range and Ly is at the peak of the filter’s transmission, the scatter is much reduced in the correlations with absolute FUV luminosity, and some possible correlations with UV color emerge, but the number of points is then too scarce for robust conclusions. Alternative explanations for the blue FUV-NUV color may include a dust effect, depressing the NUV flux. However, it is not obvious that any known interstellar extinction law would have this effect at these redshifts: the 2175Å dip, for instance, would lie at the upper (long wavelength) end of the NUV passband for redshift beyond 0.2, and the FUV would be more absorbed. There is no correlation of the H intensity with foreground E(B-V).
Figure 11 shows the magnitudes of the low-redshift sample, and their median values connected by lines. The average SED of UV-normal QSOs in this redshift range is also shown for comparison. The extended sources differ from the pointlike ones. Among pointlike sources, the overall brightness of UV-normal QSOs is slightly lower than the UV-blue QSOs. The extended UV-blue and UV-normal samples have similar SED in the optical bands, showing similarity of the host galaxy which contributes to the flux. We performed this comparison both using dereddened magnitudes, where each photometric measurement was dereddened using E(B-V) (Table 1) estimated from the Schlegel et al. (1998) maps, before averaging the sample, as well as using observed magnitudes without extinction corrections. The individual sources and the average SEDs shift correspondingly, by up to 0.4mag in UV, but the relative differences among average SEDs remain the same.
While there is no UV spectroscopy for our objects, we examined UV HST-STIS archival spectra of a sample of QSOs published by Shang et al. (2005). None of them are included in the current GALEX MIS coverage, but a few are in the GALEX AIS (which has about 10 times shorter exposures than MIS data and therefore larger photometric errors). We also computed synthetic colors for the Shang et al sample convolving the observed HST/STIS spectral fluxes with the GALEX transmission bands. We show their position on the color-color diagram (Fig. 1, small teal diamonds). A few of the Shang et al (2005) QSOs have FUV-NUV only slightly bluer than 0.1, while the rest have typical FUV-NUV colors. The QSOs in the STIS sample with FUV-NUV 0.1 have optical spectral slope and H emission similar to the standard QSO template, but most display stronger Ly and CIV, and a range of UV slopes (steeper, similar, but also flatter, than the average). The comparison, although limited to a different sample than ours and to a few bright objects, suggests that unusual UV properties may exist that cannot be predicted from the optical data.
4.2 The high redshift QSOs
The photometric SEDs of the 21 UV-blue QSOs with redshift between 1.7 and 2.6 are shown in Fig. 12, and their optical spectra in Figs. 13 and 14. The QSO with the highest redshift in the sample (z= 2.6) has an extremely red optical spectrum and appears as a very red faint object in the optical imaging (Fig. 4). It also has much larger photometric errors than the rest of the sample: FUV=22.410.19 , NUV=22.670.29, u=22.220.8, g=21.400.23, r=20.860.14. The typical QSO FUV-NUV color for its redshift is significantly redder (more than half magnitude, Fig. 1) so the object may still deserve attention.
At redshift z=2, the NUV band includes flux in the restframe range 590 - 940Å (the filter’s becomes restframe 757Å) and the FUV band includes restframe 450-590Å ( restframe 500Å). We speculate that a combination of steep flux rise towards restframe extreme ultraviolet (EUV) and absorption below the Lyman limit may explain the observed FUV-NUV color. We constructed spectral templates with FUV flux rising more steeply than in standard templates, using two power-law slopes F with =-0.6 and -1.2. The average slope between 5001200Å in the large sample of Telfer et al. (2002) is F , i.e. =-0.24 in F. Our EUV-steep templates are shown with dark green diamonds in Figs. 1 and 2. While they have synthetic FUV-NUV color bluer than the canonical template at redshift 2, they are still more than half magnitude redder than the colors observed in our UV-blue sample. The fact suggests that a combination of both EUV flux rise at shorter wavelengths and a deep absorption below the Lyman limit may be required to explain the observed colors of our UV-blue QSOs. The suggestion is supported by Fig. 6, showing the absolute NUV luminosity of our UV-blue sample to be lower than average. This can also be appreciated in Fig. 12 which shows all the observed magnitudes for our high-redshift sample. The Lyman limit in this redshift range lies between the NUV and u bands, and the Lyman drop is clearly seen. The line shows the median values, and only the object with a red optical spectrum mentioned above differs significantly (shown by the dotted line). Figure 12 also shows the median magnitudes for UV-normal QSOs from the MIS survey, with the same redshift range and error cuts. The average FUV-NUV is much ‘redder’ for the UV-normal QSOs, consistent with our selection. The UV-blue QSOs are fainter in the NUV band, which is sampling the Lyman limit at these redshifts. Thus, our QSOs sample may have somewhat more extinction and more severe absorption below the Lyman limit. Three of the UV-blue QSOs have strong BAL-type C IV absorption (Fig. 13). Their Lyman discontinuities (estimated from the broad-band photometry, as defined in Fig.15) are very large for two of them and smaller than average for one. Thus, it is not clear whether BAL absorbers contribute signifcantly to the Lyman drop.
Binette & Krongold (2008, and references therein) discuss the spectrum of Ton 34, an unusual QSO with an enhanced “Lyman valley” in its UV spectra (IUE and HST), which can be reproduced by their models of absorption from carbon crystalline dust (nanodiamonds). Ton 34 is at redshift z=1.93 and we investigated whether it could be a possible counterpart of our UV-blue QSOs. It is not included in the GALEX surveys to date (it is just outside the edge of an observed GALEX field), so we estimated GALEX FUV and NUV magnitudes by convolving the IUE SW and LW spectra of Ton 34 with the GALEX filters, and obtained FUV-NUV 1.3, close to the expected color of the UV-normal sample at this redshift and much redder than our UV-blue QSOs. This color estimate is uncertain because in the IUE spectra the signal is close to the background limit, and HST spectra of Ton 34 do not even cover one of the GALEX bands. The GALEX FUV band includes flux longwards of 1344Å, while in the IUE spectrum of Ton 34 the flux is very steeply rising just shortwards of this limit. Therefore, a slightly more redshifted analog of Ton 34 would produce a much brighter FUV magnitude.
Models of crystalline dust absorption by Binette & Krongold (2008) show that the general effect is a very deep Lyman valley, and in more detail the relative amounts of absorption in the wavelength ranges sampled by the GALEX FUV and NUV filters at z2 vary according to the dust geometry and composition. For example, comparison of dust models in figure A.2 of Binette & Krongold (2008) suggests that a lower column density of the carbon crystalline dust screen (or an intrinsic SED steeper towards short wavelengths) may produce a higher FUV flux, and small grain dust (similar composition as Milky Way dust, i.e. silicate and graphite grains, but grain sizes much smaller than MW dust and larger than nanodiamonds) would cause a significant depression of the observed-NUV flux but less reduction of the observed-FUV at the redshift of our high-z UV-blue QSOs. This effect would be qualitatively consistent with the SED of our UV-blue QSOs. From broad-band photometry alone, it is not possible to separate effects of dust absorption and intrinsic SED slope, therefore we can only speculate that the observed FUV-NUV colors in our sample are qualitatively compatible with absorption from dust with grains differing from Milky Way dust (smaller grains), and possibly a steeper flux rise towards EUV. The question remains open as to what causes the extremely blue FUV-NUV colors, and whether these objects have known counterparts with similar properties, until UV spectroscopy can be obtained.
We have measured the emission lines of C IV and C III] (EW, total flux, and full width (FW) at 10% of the peak flux above the local continuum) from the SDSS spectra. Typical errors are of the order of 10%. Ly is generally too near the end of the optical spectra and could be measured only in ten cases. The C III] line is free of absorptions but some QSOs have significant BAL and interstellar absorptions in the CIV line. The emission line properties do not correlate at all with the FUV-NUV color. The FUV-NUV color does correlate with the g-i color, which may indicate that extinction is involved, and with redshift, although weakly, in the sense that higher redshift objects are more UV-blue (Fig. 5 and 15). This is what we would expect in the rest wavelengths below the Lyman limit, where the continuum is rising again. The Lyman discontinuity is larger for higher redshifts too, which is likely caused by where it lies between the NUV and u bandpasses. The emission line EW is larger for fainter FUV magnitudes, but scales more slowly than the continuum flux. The line full-width is higher for more luminous QSOs, based on their g-band magnitudes (rest frame FUV). Flux and EW of C III] and C IV lines correlate with the Lyman discontinuity, but not the line full width. Thus, there is a connection between line emission and the EUV continuum. Figure 15 shows some of these correlations; the Spearman’s significance test gives a probability of correlation (clockwise from top left) of 99%, 94%, 58% and 99%.
While the dust absorption affects more the continuum, the ionization would be reflected by the line ratios. Binette & Krongold (2008, and references therein) discuss also the effects of shock ionization versus photoionization. It is interesting that their models show low C IV and N V relative to Ly, compared with UV-normal QSOs. We measured LyNV and C IV in our high-redshift UV QSOs where possible (10 cases). The line flux ratios may be useful diagnostic since shocks may not be related to the continuum. Therefore, we also examined SDSS spectra of UV-normal QSOs in the same redshift range and compared their line strength with the UV-blue sample. We extracted spectroscopically confirmed QSOs in the same redshift range z= 1.7-2.5, but with FUV-NUV 0.1, from our master catalog of matched sources. We found 420 objects, compared with our 21 with FUV-NUV 0.1. We imposed the same error cuts in FUV, which unfavours the red (normal) QSOs, so the ratio (5%) is a lower limit for the fraction of UV-blue QSOs compared with normal ones. The relative numbers however may be highly biased because the SDSS spectral targets were chosen with criteria not related to our UV selection. We measured the same line ratio only for the UV-normal comparison objects with z=2.2-2.5, where Ly is included in the optical spectra. The measurements are shown in Fig. 16, where a linear fit is also shown; the formal probability of correlation is over 99%. The average line ratios for the UV-blue and UV-normal samples are given in Table 4. The average is 5.2 for our UV-blue sample and 3.7 for our normal comparison sample. The collisional model predicts a ratio of 6.7, and the photoionization model 1.8. In one of our UV QSOs the ratio Ly+NV / CIV is about 10 while Ton 34 has a ratio of 8.7. This bears out the similarity with Ton 34, and a dominance of collisional ionization, compared with UV-normal QSOs.
5 Conclusions and summary
We analyzed 171 spectroscopically confirmed QSOs with FUV-NUV color bluer than 0.1, extracted from the GALEX MIS survey with complemetary SDSS optical data. Most of these objects have redshift 0.5, and we speculate that Ly emission enhanced up to a factor of 3 with respect to average templates, may explain the observed colors. Their optical properties are similar to those of UV-normal QSOs. Both photometric and emission line properties differ between point-like and extended sources, reflecting the contribution from the host galaxy in the latter. The slope of their optical spectra and the strength of H (flux and EW) correlate (increase) with intrinsic UV luminosity. Ly goes through the GALEX FUV band in the redshift range of these objects, between 0.1 and 0.5, therefore the resulting effect on the broad-band FUV magnitude is a combination of the line intensity and the filter’s transmission curve. A restricted sub-sample with redshift around 0.2 (where Ly is at the peak of the filter’s transmission) seems to show tighter correlations but it is statistically insufficient to support conclusions. The UV luminosity is brighter (0.5 to 1 mag on average) than that of our UV-normal comparison sample, the difference being larger in FUV and for the pointlike objects (Fig. 6 and 11).
Our sample of UV-blue QSOs also includes 21 objects with redshift between 1.7 and 2.6. Their photometric errors are generally large, the combined FUV-NUV 1 errors are between 0.1 and 0.37 mags, but our FUV-NUV selection limit (FUV-NUV 0.1) is bluer than typical QSO colors at this redshift by more than half magnitude, and the observed FUV-NUV colors are bluer than the typical color by up to 1 magnitude or more (Fig. 1). For these UV-blue QSOs at higher redshift we speculate that a combination of unusually strong absorption in GALEX-NUV (restframe 600-900Å) and EUV-steeply rising flux (GALEX FUV restframe 450-590Å) may explain the FUV-NUV color. This is suggested by two facts. First, ad-hoc templates with flux rising towards restframe EUV more steeply than in canonical templates, produce observed FUV-NUV colors bluer than the average template (Fig. 2), but still redder than our selection limit by 0.2 mags and redder than most of our UV-blue sample by up to 1 mag. Second, comparison with average SED of a UV-normal QSO sample, shows the NUV luminosity of the UV-blue sample to be fainter, suggesting absorption in the observed NUV. Dust with composition similar to the typical Milky Way dust but smaller grains and carbon crystalline nano-size grains (nanodiamonds) would cause absorption in the observed NUV band, according to the models of Binette and Krongold (2008), which may qualitatively account for the observed FUV-NUV colors. UV spectroscopy is needed to pinpoint the cause for the FUV-NUV color of these objects. The Ly to CIV ratio is stronger in the optical spectra of the UV-blue QSOs than in the UV-normal comparison sample (at the 95% confidence level from K-S test, although both samples are very small, see Fig. 16), suggesting collisional ionization to be more relevant in the UV-blue QSOs.
The group of UV QSOs at z2 may probe a particularly relevant phase of galaxy formation, tighly connected with the formation of the massive central black hole. In current QSO/Spheroid coevolution models (e.g. Granato et al 2004) the power of the central QSO rises almost exponentially and quickly stops the star formation process. During the previous phase it is strongly dust enshrouded and not visible except in X rays. A phase of decreasing (but still significant) extinction follows, and finally a shining phase until the fuel is consumed. A quick transition is expected between dust-extinguished and not extinguished phases for QSOs at high z. The relative percentage of UV-blue QSOs could be a measure of the relative lifetimes of that phase. As explained in section 4.2, within our sample of matched GALEX-SDSS sources, the number of spectroscopically confirmed QSOs with FUV-NUV0.1 is about 5% of those with redder FUV-NUV in the redshift range around 2. However, this number may be highly biased because the availability of SDSS spectra is serendipitous from the point of view of our selection. The fraction of sources with available spectra is not uniform across the range of optical and UV colors, and redshift, of our photometric candidate sample. Spectroscopic selection especially favours the brightest samples, while Bianchi et al. (2007, 2008) show for example a steep increase of UV-blue extragalactic object candidates at faint magnitudes. Some UV-to-optical color ranges are also contaminated by stellar objects and the purity of photometric candidate samples varies greatly according to the colors regime and parameters. The aim of this work was to point out that a non-negligible sample of UV-blue QSOs exist, and explore their nature. Statistical considerations will be addressed using a larger sample.
Another possible bias may arise from variability, which is frequently observed in QSOs. Serendipitous repeated UV observations for our sample show variations by 3 in some objects, and in many cases the FUV-NUV in repeated measurements is redder than in our selected data-set, making some of these objects UV-normal or close to normal in some of the measurements, and extremely blue in others. We have tried to exclude as thoroughly as possible imaging or pipeline artifacts, using the flags provided by the pipeline and identifying several additional unreliable measurements by individual analysis. However, we should keep in mind that pipeline photometry of large datasets has statistical value, and in particular the combination of GALEX and SDSS source catalogs over a large area of the sky proved invaluable to characterize elusive classes of objects (Bianchi et al. 2007), but it should not be overinterpreted for individual objects.
For both the low-redshift QSOs where Ly may be stronger (possibly up to 3) than in typical QSO templates, and the high-redshift QSOs where deep Lyman valley absorption may occur, UV spectroscopy is needed for a conclusive explanation of the FUV-NUV color, and to assess whether these are similar to some known objects. Our analysis showed that the GALEX photometry provides a unique sieve to select these UV-blue QSOs, whose optical properties are not unusual. The analysis of this limited spectroscopic sample, and its average photometric properties, also provides useful information to separate these QSOs from stellar binaries with a hot WD in our larger samples of photometric candidates (e.g. Bianchi et al. 2007, 2008), as shown in Figs 1 and 2. The contamination of these objects in stellar samples may be very significant at faint magnitudes, because the density of Milky Way hot WDs, extracted from GALEX catalogs, at MIS depth (UV mag 22.7 ABmag) is much lower than that of QSOs (Bianchi e al. 2007, 2008).
- (paper I) Bianchi,L., 2008, in ”Space Astronomy: The UV Window to the Universe”, APSS, DOI: DOI: 10.1007/s10509-008-9761-3 (also available at http://dolomiti.pha.jhu.edu/publgoto.html)
- () Bianchi,L., et al. 2009, in preparation
- () Bianchi,L., Efremova, B., Herald, J. et al. 2008, in ”Future Directions in Ultraviolet Spectroscopy”, eds. Van Steenberg et al., in press
- () Bianchi,L., et al. 2007,ApJS, 173, 659
- () Bianchi, L., et al. 2006, in “UV Astronomy: Stars from Birth to Death”, eds. A.I. Gomez de Castro and M. Barstow, UCM Editorial Complutense, p. 95
- () Bianchi,L., et al. 2005,ApJ, 619, L27
- () Binette, L. & Krongold, Y., 2008, A&A, 478, 739
- () Granato, G.L. et al. 2005, ApJ, 600, 580
- () Hutchings, J.B., & Bianchi, L. 2008, PASP, 120, 275
- () Hutchings, J.B., Scholz, P. & Bianchi, L. 2009, AJ, submitted
- () Martin, C., et al. 2005, ApJ, 619, L1
- () Morrissey, P., et al. 2007, ApJS, 173, 682
- () Telfer, R. et al. 2002, ApJ 565, 773
- () Shang, Z. et al 2005, Apj, 619, 41
- () Schlegel, D.J., Finkbeiner, D.P., & Davis, M. 1998, ApJ, 500, 525
- () York, D. et al. 2000, AJ, 120, 1579
- () Zheng, W., et al. 1997, ApJ, 475, 469
|GALEX IAU ID||RA (deg)||Dec(deg)||FUV (AB)||NUV (AB)||u (AB)||g (AB)||r (AB)||i (AB)||z (AB)||redshift||E(B-V)||Comment|
|High Redshift QSOs (z1)|
|GALEX J075627.72+205415.1||119.1154984||20.9041836||21.00 0.15||21.82 0.14||20.23 0.09||19.72 0.02||19.52 0.03||19.22 0.04||19.22 0.15||2.21||0.06||pointlike|
|GALEX J025221.08-085516.0||43.0878540||-8.9211193||20.88 0.09||21.67 0.11||18.56 0.03||18.03 0.01||17.94 0.01||17.85 0.02||17.50 0.04||2.29||0.05||pointlike|
|GALEX J161821.57+492603.0||244.5898736||49.4341532||22.05 0.18||22.74 0.23||19.23 0.04||18.64 0.01||18.66 0.01||18.63 0.02||18.38 0.05||2.28||0.02||pointlike|
|GALEX J222323.21-084740.5||335.8467089||-8.7945764||22.35 0.14||23.01 0.25||19.52 0.05||19.47 0.02||19.38 0.02||19.14 0.03||19.00 0.10||1.99||0.05||pointlike|
|GALEX J145645.30+595436.2||224.1887297||59.9100553||22.57 0.19||23.14 0.29||20.52 0.08||20.31 0.03||20.41 0.04||20.44 0.07||19.97 0.15||2.17||0.01||pointlike|
|GALEX J211858.56-063027.3||319.7439988||-6.5075842||22.17 0.17||22.58 0.19||19.97 0.06||19.75 0.02||19.62 0.02||19.48 0.04||19.27 0.11||2.00||0.11||pointlike|
|GALEX J082337.98+455621.4||125.9082468||45.9392840||22.41 0.19||22.67 0.29||22.22 0.80||21.40 0.23||20.86 0.14||20.33 0.09||19.46 0.11||2.59||0.04||pointlike|
|GALEX J003505.58-103536.4||8.7732425||-10.5934493||22.46 0.23||22.68 0.25||19.35 0.04||18.57 0.01||18.62 0.01||18.49 0.01||18.20 0.05||2.26||0.03||pointlike|
|GALEX J230724.98+000958.0||346.8540794||0.1661096||21.93 0.15||22.12 0.10||19.40 0.06||19.41 0.02||19.33 0.03||19.04 0.03||19.23 0.17||1.90||0.04||pointlike|
|GALEX J113638.68+011031.5||174.1611550||1.1754105||21.94 0.18||22.13 0.20||19.83 0.05||19.68 0.02||19.60 0.03||19.47 0.04||19.17 0.11||2.13||0.02||pointlike|
|GALEX J155013.52+033744.6||237.5563540||3.6290531||22.74 0.25||22.86 0.27||19.54 0.05||19.51 0.02||19.33 0.02||18.93 0.03||18.76 0.09||1.91||0.10||pointlike|
|GALEX J233139.76+010428.7||352.9156781||1.0746313||20.56 0.07||20.67 0.07||18.93 0.03||18.54 0.01||18.45 0.01||18.53 0.01||18.40 0.05||2.25||0.04||pointlike|
|GALEX J090814.51+550701.6||137.0604696||55.1171069||22.98 0.17||23.03 0.22||20.21 0.06||19.84 0.02||19.46 0.02||19.07 0.02||18.90 0.08||1.93||0.02||pointlike|
|GALEX J015229.22-083640.0||28.1217418||-8.6111176||22.43 0.19||22.47 0.19||20.27 0.10||19.84 0.03||19.27 0.03||18.84 0.03||18.83 0.11||1.72||0.02||pointlike|
|GALEX J020419.29+143929.2||31.0803956||14.6581219||21.07 0.11||21.11 0.09||19.11 0.03||18.85 0.01||18.80 0.01||18.61 0.02||18.43 0.06||2.17||0.05||pointlike|
|GALEX J082616.05+502049.5||126.5668879||50.3470827||22.58 0.20||22.61 0.27||19.19 0.04||18.75 0.01||18.62 0.01||18.60 0.02||18.37 0.05||2.19||0.04||pointlike|
|GALEX J005408.42-094637.1||13.5350664||-9.7769584||22.64 0.28||22.68 0.24||18.45 0.02||18.06 0.01||17.82 0.01||17.61 0.01||17.41 0.03||2.13||0.04||pointlike|
|GALEX J145055.62+015419.0||222.7317516||1.9052814||21.48 0.12||21.45 0.12||19.86 0.05||19.77 0.02||19.75 0.03||19.56 0.04||19.63 0.18||1.87||0.04||pointlike|
|GALEX J141807.07+050431.3||214.5294395||5.0753685||20.92 0.09||20.87 0.06||19.93 0.06||19.56 0.02||19.10 0.02||18.72 0.02||18.46 0.05||1.89||0.03||pointlike|
|GALEX J085344.92+565337.9||133.4371727||56.8938634||21.71 0.13||21.63 0.11||20.42 0.14||20.45 0.05||20.40 0.08||20.12 0.10||21.35 1.50||1.86||0.05||pointlike|
|GALEX J084658.13+465011.4||131.7422222||46.8364994||21.75 0.08||21.66 0.08||19.47 0.04||19.36 0.02||19.24 0.02||19.14 0.03||18.88 0.07||2.12||0.02||pointlike|
|Low Redshift QSOs (z 1)|
|GALEX J075304.64+252436.6||118.2693129||25.4101641||19.78 0.08||20.67 0.06||19.36 0.07||18.84 0.02||18.33 0.02||17.94 0.02||17.72 0.06||0.15||0.09||extended|
|GALEX J154912.35+030641.8||237.3014425||3.1116111||22.45 0.28||22.95 0.28||21.62 0.57||20.34 0.06||19.01 0.03||18.48 0.03||17.96 0.09||0.25||0.13||extended|
|GALEX J140816.07+015528.5||212.0669475||1.9245728||22.26 0.18||22.73 0.29||20.50 0.28||19.11 0.03||18.06 0.02||17.61 0.02||17.23 0.05||0.17||0.03||extended|
|GALEX IAU ID||FUV(AB)||NUV(AB)||Date||exp.time (s)||matched GALEX ID||Dist(”)||FUV(AB)||NUV(AB)||Date||exp.time(s)||Survey||Comments|
|High Redshift QSOs (z1)|
|J230724.98+000958.0||21.930.15||22.120.10||9/9/2004||1697 /3055||J230724.98+000958.4||0.44||21.610.10||21.450.08||8/24/2003||3181 /3181||MIS||!|
|J005408.42-094637.1||22.640.28||22.680.24||9/23/2003||1666 /1666||J005408.44-094637.7||0.71||23.140.38||21.900.14||9/25/2004||1648 /1648||GII|
|J085344.92+565337.9||21.710.13||21.630.11||1/17/2004||1694 /1694||J085344.84+565338.6||1.00||22.050.10||21.600.09||1/17/2004||2959 /2959||MIS||e E|
|Low Redshift QSOs (z1)|
|J075304.64+252436.6||19.790.08||20.670.06||2/14/2006||1703 /1703||J075304.67+252436.6||0.53||20.450.06||2/15/2006||1698 /1698||MIS||epb E|
|J153219.90+033811.1||20.920.13||21.200.12||6/7/2003||828 /828||J153219.90+033812.3||1.21||21.450.19||20.910.13||6/7/2003||476 /476||MIS|
|J223553.88+142805.7||19.780.04||19.970.03||8/13/2005||1608 /3007||J223553.87+142806.0||0.35||19.760.01||19.640.01||8/23/2003||31533/31533||DIS||! EB|
|J160655.42+534016.9||18.980.02||19.110.02||6/25/2004||2862 /2862||J160655.40+534016.7||0.29||19.710.22||19.420.01||5/2/2005||42 /13341||DIS||! b B|
|J085318.52+551525.3||21.880.14||21.980.14||1/18/2004||1694 /1694||J085318.57+551525.3||0.48||22.430.19||22.060.14||1/3/2006||1700 /1700||GII|
|J014248.83+142126.9||21.690.14||21.780.16||10/4/2004||1630 /1630||J014248.85+142126.1||0.86||22.240.16||21.940.11||11/10/2004||1702 /3391||GII|
|J040446.72-045429.7||21.180.10||21.240.09||11/3/2004||2462 /2462||J040446.73-045431.3||1.60||21.170.12||20.580.11||11/3/2004||1535 /1535||MIS||! eb EPB|
|J013928.93-103425.9||20.810.08||20.850.06||10/15/2003||1579 /1579||J013928.89-103426.5||0.77||20.590.05||20.580.04||12/13/2004||3284 /3753||GII||!|
|J234019.81+005907.9||19.180.02||19.200.02||8/23/2003||3145 /3145||J234019.83+005910.3||2.46||19.450.04||19.190.03||9/22/2006||1669 /1669||MIS||! EP|
|J235457.10+004220.5||18.250.03||18.220.02||10/22/2006||1108 /1108||J235457.15+004220.3||0.72||18.340.03||18.200.02||10/21/2006||924 /924||MIS||EB|
|J160545.93+532209.9||18.260.02||18.220.01||7/31/2004||1637 /1637||J160545.98+532210.8||1.00||18.080.12||17.860.00||5/2/2005||42 /13341||DIS||! PB|
|J032225.43-081255.3||18.930.03||18.880.02||11/29/2003||1698 /1698||J032225.37-081255.4||0.90||18.920.04||18.720.02||11/30/2003||1698 /1698||MIS||! e E|
|J160815.26+524450.8||19.460.04||19.390.03||7/31/2004||1635 /1635||J160815.21+524451.3||0.71||19.860.28||19.730.01||5/3/2005||40 /14866||DIS||! e EPB|
|J161156.36+521116.9||19.170.04||19.090.02||7/31/2004||1635 /1635||J161156.35+521117.2||0.36||19.100.04||19.000.02||7/31/2004||1631 /1631||NGS||! eb B|
|J005057.44+143753.7||21.340.11||21.260.10||9/25/2003||1687 /1687||J005057.45+143753.8||0.21||21.700.11||21.400.09||8/30/2003||1967 /1967||MIS|
|J005328.80-085754.8||18.440.02||18.360.01||9/16/2003||1602 /1602||J005328.79-085754.1||0.70||18.330.02||18.220.01||9/23/2003||1666 /1666||MIS||!|
|J163625.47+421346.1||18.440.02||18.350.02||5/23/2004||1702 /1702||J163625.48+421346.8||0.72||18.310.03||18.260.01||7/8/2004||1195 /1586||DIS||!|
|J233633.71-092616.1||18.440.02||18.340.01||9/2/2006||2189 /2189||J233633.72-092616.9||0.82||18.420.02||18.240.01||9/27/2005||1973 /2859||MIS||! eb EB|
Note. – Explanation of comments: ”!” denotes magnitude discrepancy (NUV or FUV) greater than 3, ”E” denotes that the EDGE artifact flag is set, ”P” denotes SExtractor flag 1 set (object has neighbours) ”B” denotes SExtractor flag 2 set (object was originally blended with another one). Lower case flag codes for original sources, upper case for matched measurements.
|GALEX IAU ID||H Width||EW||log flux|
|[Å]||[Å]||10 ergs cm sÅ|
|UV QSOs||UV-Normal QSOs|
|Ly flux||440||285 (365)|
|C IV flux||90||75 (96)|
|C III] flux||52||41 (52)|
|C IV/ C III]||1.74||1.82|
Note. – Values in parentheses for UV-normal QSOs are continuum-corrected. Note that the average of the ratios and the ratio of the averages are not the same because the average fluxes weigh the high values more. Theoretical values for Ly/C IV ratio are 6.7 for collisional ionization and 1.8 for photoionization.