The Detection of a Population of SubmillimeterBright, Strongly-Lensed Galaxies
Department of Physics and Astronomy, The Open University, Walton Hall, Milton Keynes, MK7 6AA, UK
Istitute Nazionale di Astrofisica (INAF), Osservatorio Astronomico di Padova, Vicolo Osservatorio 5, I-35122 Padova, Italy
Scuola Internazionale Superiore di Studi Avanzati (SISSA), Via Bonomea 265, I-34136 Trieste, Italy
Department of Physics and Astronomy, University of California, Irvine, CA 92697, USA
Oxford Astrophysics, Denys Wilkinson Building, University of Oxford, Keble Road, Oxford, OX1 3RH, uk
Jet Propulsion Laboratory (JPL), Pasadena CA, 91009, USA
California Institute of Technology, Pasadena, CA 91125, USA
National Radio Astronomy Observatory Post Office Box 2, Green Bank, WV 24944, USA
Harvard-Smithsonian Center for Astrophysics, Cambridge, MA 02138, USA
Institut d’Astrophysique de Paris, Universitte Pierre et Marie Curie and CNRS, 98 bis boulevard Arago, 75014 Paris, France
Institut de Radioastronomie Millimétrique (IRAM), 300 rue de la piscine, 38406 Saint-Martin d’Hères, France
Laboratoire Astrophysique, Instrumentation et Modlisation Paris Saclay, Commissariat l’nergie Atomique (CEA)/Direction des Sciences de la Matire (DSM) - CNRS - Universit Paris Diderot, Institut de recherche sur les lois fondamentales de l’Univers (Irfu)/Service d’Astrophysique, CEA Saclay, Orme des Merisiers, F-91191 Gif-sur-Yvette Cedex, France
Physics Department, University of Johannesburg, Post Office Box 524, Auckland Park 2006, South Africa
SETI Institute, 515 North Whisman Avenue Mountain View CA, 94043, USA
Department of Physics, University of Crete, 71003 Heraklion, Greece
Dipartimento di Astronomia, Universit di Padova, Vicolo Osservatorio 2, I-35122 Padova, Italy
Centre for Astrophysics Research, Science and Technology Research Institute, University of Hertfordshire, Herts AL10 9AB, UK
UK Astronomy Technology Center, Royal Observatory Edinburgh, Edinburgh, EH9 3HJ, UK
Scottish Universities Physics Alliance, Institute for Astronomy, University of Edinburgh, Royal Observatory,Edinburgh, EH9 3HJ, UK
School of Physics and Astronomy, Cardiff University, The Parade, Cardiff, CF24 3AA, UK
Instituto Nacional de Astrofísica, Óptica y Electrónica, Apartado Postal 51 y 216, 72000 Puebla, Mexico
Argelander Institut fur Astronomie, Universit Bonn, Auf dem Hgel 71, 53121 Bonn, Germany
Herschel Science Centre, European Space Astronomy Centre, European Space Agency (ESA), Post Office Box 78, 28691 Villanueva de la Caada, Madrid, Spain
School of Physics and Astronomy, University of Nottingham, University Park, Nottingham NG7 2RD, UK
Scottish Universities Physics Alliance, School of Physics and Astronomy, University of St. Andrews, North Haugh, St. Andrews, KY16 9SS, UK
Instituto de Astrofísica de Canarias, C/Vía Láctea s/n, E-38200 La Laguna, Spain
Departamento de Astrofísica, Universidad de La Laguna (ULL), E-38205 La Laguna, Tenerife, Spain
Eureka Scientific, 2452 Delmer Street, Suite 100, Oakland CA 94602, USA
Goddard Space Flight Center, Code 667, Greenbelt Road, Greenbelt, MD 20771, USA
Astrophysics Group, Physics Department, Blackett Lab, Imperial College London, Prince Consort Road, London SW7 2AZ, UK
Sterrenkundig Observatorium, Universiteit Gent, Krijgslaan 281 S9, B-9000 Gent, Belgium
Department of Physics and Astronomy, University of Louisville, Louisville, KY 40292, USA
Institut d’Astrophysique Spatiale (IAS), Btiment 121, F-91405 Orsay, France
Universit Paris-Sud 11 and CNRS (UMR 8617), 91400 Orsay, France
Instituto de Fisica de Cantabria, Consejo Superior de Investigaciones Cientficas - Universidad de Cantabria, Avenue de Los Castros s/n, Santander, 39005, Spain
Astrophysics Research Institute, Liverpool John Moores University Twelve Quays House, Egerton Wharf, Birkenhead, CH41 1LD, UK
Astrophysics Branch, NASA Ames Research Center, Mail Stop 245-6, Moffett Field, CA 94035, USA
NASA Goddard Space Flight Center, Code 667, Greenbelt Road, Greenbelt, MD 20771, USA
Department of Astrophysics and The Enrico Fermi Institute, University of Chicago, 5640 South Ellis Avenue, Chicago, IL 60637, USA
Department of Physics and Astronomy, University of Pennsylvania, Philadephia, PA 19104, US
Rutgers, the State University of New Jersey, Department of Physics and Astronomy, 136 Frelinghuysen Road, Piscataway, NJ 08854-8019, USA
HH Wills Physics Laboratory, University of Bristol, Tyndall Avenue, Bristol BS8 1TL, UK
Laboratoire d’Astrophysique de Marseille and Aix-Marseille Universit, UMR6110 CNRS, 38 rue F. Joliot-Curie, F-13388 Marseille, France
School of Mathematical Sciences, Queen Mary, University of London, Mile End Road, London, E1 4NS, UK
University of Colorado, Center for Astrophysics and Space Astronomy, 389-UCB, Boulder, CO 80303, USA
Department of Astronomy, University of Maryland, College Park, MD 20742 USA
Observatories of the Carnegie Institution, 813 Santa Barbara Street, Pasadena, CA 91101, USA
Institute for Space and Astronautical Science, Japan Aerospace Exploration Agency, 3-1-1 Yoshinodai, Chuo-ku, Sagamihara 252-5210 Japan
Infrared Processing and Analysis Center, Pasadena, CA 91125, USA
Jeremiah Horrocks Institute, University of Central Lancashire, Preston PR1 2HE, UK
Space Science and Technology Department, Rutherford Appleton Laboratory, Chilton, Didcot, Oxfordshire OX11 0QX, UK
University of British Colombia, 6224 Agricultural Road, Vancouver, BC V6T 1Z1, Canada
Institute for Computational Cosmology, Durham University, Durham DH1 3LE, UK
Max Planck Institut fuer Kernphysik (MPIK), Saupfercheckweg 1, 69117 Heidelberg, Germany
Leiden Observatory, Leiden University, Post Office Box 9513, NL - 2300 RA Leiden, The Netherlands
to whom correspondence should be addressed E-mail: firstname.lastname@example.org
Gravitational lensing is a powerful astrophysical and cosmological probe and is particularly valuable at submillimeter wavelengths for the study of the statistical and individual properties of dusty starforming galaxies. However the identification of gravitational lenses is often time-intensive, involving the sifting of large volumes of imaging or spectroscopic data to find few candidates. We used early data from the Herschel Astrophysical Terahertz Large Area Survey to demonstrate that wide-area submillimeter surveys can simply and easily detect strong gravitational lensing events, with close to 100 efficiency.
When the light from a distant galaxy is deflected by a foreground mass commonly a massive elliptical galaxy or galaxy cluster or group its angular size and brightness are increased, and multiple images of the same source may form. This phenomenon is commonly known as gravitational lensing  and can be exploited in the study of high-redshift galaxy structures down to scales difficult (if not impossible) to probe with the largest telescopes at present [2, 3, 4] and to detect intrinsically faint objects. Surveys conducted at submillimeter wavelengths can particularly benefit from gravitational lensing because submillimeter telescopes have limited spatial resolution and consequently high source confusion, which makes it difficult to directly probe the populations responsible for the bulk of background submillimeter emission [5, 6]. In addition, galaxies detected in blank-field submillimeter surveys generally suffer severe dust obscuration and are therefore challenging to detect and study at optical and near infra-red (NIR) wavelengths. By alleviating the photon starvation, gravitational lensing facilitates follow-up observations of galaxies obscured by dust and in particular the determination of their redshift . Previous submillimeter searches for highly magnified background galaxies have predominantly targeted galaxy cluster fields . In fact, a blind search for submillimeter lensing events requires large area because of their rarity, and sub-arcseconds angular resolutions to reveal multiple images of the same background galaxies. Although the first requirement has recently been fulfilled, thanks to the advent of the South Pole Telescope (SPT)  and the Herschel Space Observatory (Herschel) , the second is still the prerogative of ground-based interferometric facilities, such as the Submillimeter Array (SMA) and the IRAM Plateau de Bure Interferometer (PdBI), which because of their small instantaneous field of view are aimed at follow-up observations rather than large-area survey campaigns. Nevertheless, several authors [11, 12, 13, 14] have suggested that a simple selection in flux density, rather than surveys for multiply-imaged sources, can be used to easily and efficiently select samples of strongly gravitationally-lensed galaxies in wide-area submillimeter and millimeter surveys. The explanation for this lies in the steepness of the number counts (the number of galaxies at a given brightness) of dust-obscured star-forming galaxies, which are usually referred to as submillimeter galaxies (SMGs) . Because of that, even a small number of highly-magnified SMGs can substantially affect the shape of the bright end of the submillimeter source counts enhancing the number of SMGs seen at bright flux densities than would be expected on the basis of our knowledge of the un-lensed SMG population (Fig.1). Furthermore, the frequency of lensing events is relatively high in the submillimeter  because SMGs are typically at high redshift () , and this increases the probability that a SMG is in alignment with, and therefore lensed by, a foreground galaxy. Other important contributors to the bright tail of the submillimeter counts are low-redshift () spiral and starburst galaxies  and higher redshift radio-bright Active Galactic Nuclei (AGNs) ; however both of these are easily identified, and therefore removed, in relatively shallow optical and radio surveys. Therefore, flux-density limited submillimeter surveys could provide a sample of lens candidates from which contaminants can be readily removed, leaving a high fraction (close to 100) of gravitational lens systems (Fig.1). Because this selection of lens candidates relies only on the properties of the background source (its flux density), it can probe a wide range of lens properties (such as redshifts and masses) and thus provide a valuable sample for studying the elliptical properties of lensing galaxies  as well as investigating the detailed properties of the lensed SMGs.
The submillimeter lens candidate selection at work.
Although the approach presented above may be more efficient and vastly
more time-effective than those exploited so far in the radio
 or the optical [21, 22], at least
several tens of square degrees (deg) of the sky must be observed
in the submillimeter to produce a statistically significant sample of
strongly lensed objects and a minimal contamination from unlensed
galaxies. This is because the surface density of lensed submillimeter
galaxies is predicted to be lower than 0.5 deg, for flux
densities above 100mJy at 500m (Fig.1). Submillimeter
surveys conducted before the advent of Herschel were either limited to
small areas of the sky [15, 23], or were severely
affected by source confusion due to poor spatial
resolution. Therefore no previous test of this
selection method has been performed, although the SPT has recently
mapped an area of more than 80 deg at millimetre wavelengths
 and found an
“excess” of sources that could be accounted for by a population of gravitationally-lensed objects.
The Herschel Astrophysical Terahertz Large Area Survey (H-ATLAS)  represents the largest-area submillimeter survey being currently undertaken by Herschel. H-ATLAS uses the Spectral and Photometric Imaging REceiver (SPIRE) [26, 27] and the Photodetector Array Camera and Spectrometer (PACS) [28, 29] instruments and, when completed, will cover 550deg of the sky from 100 to 500m. H-ATLAS has been designed to observe areas of the sky with previously existing multi-wavelength data: Galaxy Evolution Explorer (GALEX) ultra violet (UV) data, Sloan Digital Sky Survey (SDSS) optical imaging and spectroscopy, NIR data from the UK Infrared Telescope (UKIRT) Infrared Deep Sky Survey (UKIDSS) Large Area Survey (LAS), spectra from the Galaxy And Mass Assembly GAMA  project, radio imaging data from the Faint Images of the Radio Sky at Twenty-cm (FIRST) survey and the NRAO Very Large Array Sky Survey (NVSS). The first 14.4deg of the survey, centred on J2000 RA 09:05:30.0 DEC 00:30:00.0 and covering 3 of the total area, was observed in November 2009 as part of the Herschel Science Demonstration Phase (SDP). The results were a catalog of 6600 sources  with a significance , in at least one SPIRE waveband, where the noise, , includes both instrumental and source confusion noise and corresponds to 7 to 9 mJy/beam.
The Herschel/SPIRE 500m channel is favourable for selecting lens candidates, because the submillimeter source counts steepen at longer wavelengths [24, 32]. We used theoretical predictions  to calculate the optimal limiting flux density, above which it is straightforward to remove contaminants from the parent sample and maximize the number of strongly lensed high-redshift galaxies. The surface-density of un-lensed SMGs is predicted to reach zero by mJy  and these objects are only detectable above this threshold if gravitationally lensed by a foreground galaxy (Fig.1). The H-ATLAS SDP catalog contains 11 sources with 500m flux density above 100mJy. Ancillary data in the field revealed that six of these objects are contaminants; four are spiral galaxies with spectroscopic redshifts in the range of 0.01 to 0.05 [see  for a detailed analysis of one of these sources], one is an extended Galactic star forming region, and one is a previously known radiobright AGN . Although the number of these sources are few at bright flux densities, the measured surface densities are consistent with expectations (Fig.1) [17, 18]. Exclusion of these contaminants left the five objects that form our sample of lens candidates (tableS1) , identified as ID9, ID11, ID17, ID81 and ID130.
Unveiling the nature of the lens candidates.
For gravitational lensing systems selected at submillimeter
wavelengths, we would expect the lensing galaxy to be seen in optical
and/or NIR images, in which the emission from the lens dominates over the
higher redshift background SMG. In line with these expectations, all
of the lens candidates have a close counterpart in SDSS or UKIDSS
images (or both). A likelihood ratio analysis  showed
that the probability of a random association between these bright
submillimeter sources and the close optical/NIR counterparts is less
than a few per cent. Therefore the optical and submillimeter emissions
must be physically related, either because they occur within the same
object or because of the effects of gravitational lensing, boosting
the flux of the background source and indirectly affecting the
likelihood ratio calculations. The redshift measurements support the
later scenario. Although the optical/NIR
photometric/spectroscopic redshifts lie in the range of to 0.9 (table 1 and figs S3 and S4) , the redshifts estimated from
the submillimeter/millimeter spectral energy distributions (SED)
(table2) [following the method described in
[37, 38]], are distinctly different (Table 1). The
lensed SMG photometric redshifts have been confirmed and made more
precise through the spectroscopic detection, in these objects, of
carbon monoxide (CO) rotational line emission which are tracers of
molecular gas assocatiated to star forming environments. Until
recently, these kind of detections were difficult to achieve without
prior knowledge of the source redshift, which required extensive
optical/NIR/radio follow-up observations. Because of the development of
wide-bandwidth radio spectrometers capable of detecting CO lines over
a wide range of redshifts, it is now possible for blind redshift
measurements of SMGs to be taken without relying on optical or NIR
spectroscopy [39, 40]. ID81 was observed with the Z-Spec
spectrometer [41, 42] on the California Institute of Technology Submillimeter
Observatory (CSO). The data revealed several CO lines redshifted into the
frequency range of 187 to 310GHz; the strongest of these lines has
been interpreted as the CO J=76 line, with an estimated redshift of
. This represents the first blind redshift
determination by means of Z-Spec. We followed up this observation with the PdBI
and detected CO J=32 and CO J=54 emission lines, redshifted to
, confirming the Z-Spec measured redshift . We
also used the Zpectrometer instrument [44, 45] on the
NRAO Robert C. Byrd Green Bank Telescope (GBT) to obtain an
independent confirmation of the redshift of ID81 (table1 and fig.S1) [46, 35] and to measure the redshift of ID130.
We detected redshifted CO J=10 emission at in the
spectrum of ID130 (fig.S1) [46, 35]]. This
redshift was confirmed by the PdBI with the observation of CO J=3-2
and CO J=5-4 lines, yielding a redshift of . The
Z-Spec spectrometer observed the remaining three lens candidates
 and detected CO lines at redshifts of and
for ID9 (fig.S2)  and ID11, respectively, which
are higher and inconsistent with the redshifts derived from the
optical photometry/spectroscopy (Table1). The Z-Spec CO
measurements for ID17 are indicative of two redshifts; one, ,
that is in agreement with the optical redshift and a higher one,
, which is indicative of a more distant
To determine the morphological type of the foreground galaxies we obtained high resolution optical images for all five objects with the Keck telescope at - and -bands . ID9, ID11, ID81 and ID130 all have optical profiles that are consistent with elliptical galaxies (figs S5 and S6 and table S4) . The interpretation of the results for ID17 is complicated by the presence of two partially superimposed galaxies in the optical images (fig.S7) , neither exhibiting the disturbed morphology expected for lensed objects. This indicates that ID17 may be a gravitational lens system with two foreground lensing masses at similar redshifts ( to 0.9) possibly a merging system with some molecular gas responsible for the CO emission detected by Z-Spec at and confirmed with optical spectroscopy (table 1). A fit to the UV/optical/NIR SEDs of ID9, ID11, ID81 and ID130 , using the models of , gives stellar masses in the range of to (Table 2) and almost negligible present-day star formation, which is consistent with elliptical galaxies (fig.2).
For all five lens systems the background source appears to be undetected in the Keck - and -band images, despite the flux magnification due to lensing. After subtracting the best fit light profile from each lens we found no structure that could be associated with the background source in the residual images (figsS5 and S6) . We derived 3- upper limits from the residual maps (tableS4)  and corresponding NIR limits from the UKIDSS images. These upper limits were used to fit the SEDs of the background sources assuming the models of , calibrated to reproduce the UV-to-infrared SEDs of local, purely star-forming ultra luminous infrared galaxies (ULIRG; 10L/L10)  (fig.2). A visual extinction  of 2 is required to be consistent with the optical/NIR upper limits (fig.2 and table2), confirming severe dust obscuration in these galaxies along the line-of-sight. Our results indicate that these submillimeter bright gravitationally lensed galaxies would have been entirely missed by standard optical methods of selection.
We obtained observations at the SMA for ID81 and ID130 at 880m, with the aim of detecting the lensed morphology of the background galaxy . The SMA images reveal extended submillimeter emission distributed around the cores of the foreground elliptical galaxies, with multiple peaks (four main peaks in ID81 and two in ID130), which is consistent with a lensing interpretation of these structures (Fig. 3). The position of these peaks can be used to directly constrain the Einstein radius the radius of the circular region on the sky (the Einstein ring) into which an extended source would be lensed if a foreground galaxy were exactly along the line of sight of the observer to the source (for a perfectly circular lens). The Einstein radius is a measure of the projected mass of the lens, so it can be used to derive the total (dark plus luminous) mass of the galaxy within the Einstein radius (table2) . Another measure of the total mass of a lens is the line-of-sight stellar velocity dispersion, . We have estimated from the local FaberJackson (FJ) relation  between and the rest-frame B-band luminosity for elliptical galaxies. Assuming passive stellar evolution for the lens galaxies, which is appropriate for elliptical galaxies, we have extrapolated their rest-frame K-band luminosity to [using the evolutionary tracks of ], and then converted this to B-band luminosity using the B-K=4.43 color relation from . The result was then applied to the FJ relation from . Given a mass model for the lens , we can predict the Einstein radius of the galaxy from the value of expected from the FJ relation and compare it with that directly measured from the SMA images (Table 2). Although the value of the Einstein radius derived from the line-of-sight stellar velocity dispersion is affected by large uncertainties (as a result of the scatter in the FJ relation) it is consistent with the value measured in the SMA images for both ID81 and ID130. In order to test whether the properties of the lensing galaxies in our sample are consistent with those of other known lens ellipticals at similar redshift, we compared the V-band mass-to-light ratio of the lens galaxy for ID81 and ID130 (Table2)  to those measured in the Sloan Lens Advanced Camera for Surveys (ACS) (fig.4) , which cover a similar redshift range (z0.1 to 0.3). The agreement with the average trend revealed by SLACS confirms that our lens selection method is not biased to lensing ellipticals with atypical luminosities. Moreover, the location of ID130 in Fig.4 indicates that our selection method can probe lower masses and lower luminosity lens galaxies than those sampled by SLACS, thus offering a wider range in lens properties to be investigated.
The best fit SED to the submillimeter/millimeter photometry for each of the five background sources give infrared luminosities L310L (Table2), which would classify these objects as Hyper Luminous Infra-Red galaxies (HLIRGs; L10L). However, a correction for magnification because of lensing will reduce these values by a factor of 10 or greater. For example, assuming that the light distribution of the background source is described by a Gaussian profile with a full width at half maximum (FWHM) of 0.2arcseconds [which is consistent with the physical extension of the background galaxy in ], the best-fit lens model (fig.S9)  predicts a total amplifications of 19 and 6 for ID81 and ID130, respectively. Typical amplifications of 8 to 10 are also suggested by , therefore, it is more likely that these sources are ULIRGs.
These results already provide constraints for models of the formation and evolution of massive galaxies at high redshift. The fact that many (if not all) of the brightest SMGs detected in the H-ATLAS SDP field are amplified by lensing, implies that un-lensed star-forming galaxies with flux densities more than 100mJy at 500m are rare, with of them per 14.4deg, at 99 probability (assuming Poisson statistics). This translates into a 0.32deg upper limit on the surface density of these sources. The same limit should translate to the abundance of HLIRGs with L510Lat , because they would also have 500-m flux densities above mJy, which has possible implications for the role of feedback during the formation of the most massive galaxies in the universe. By extrapolating our SDP findings to the full H-ATLAS field, we predict a total sample of more than 100 bright lensed sources, with which we can further improve this constraint.
|17||0.770.13||0.94350.0009||2.0||0.9420.004 & 2.3080.011|
|Datum is from CSO/Z-Spec |
|Datum is from the William Herschel Telescope |
|Datum is from SDSS|
|Datum is from GBT/Zpectrometer |
|Datum is from PdBI |
|Datum is from the Apache Point Observatory |
Supporting Online Material
The five H-ATLAS/SDP lens candidates, i.e. ID9, ID11, ID17, ID81 and ID130, were observed on 2010 March 9 and 10, with MAMBO, at the IRAM 30 meter telescope on Pico Veleta, in Diretcor’s Discretionary Time (DDT). The MAMBO array consists of 117 bolometer elements and operates at a central frequency of 250GHz, corresponding to 1.2mm. The beam size (FWHM11 arcsec) of MAMBO ensures that the true dust emission at 1.2mm is obtained if the source is not more extended than a few arcseconds. Each science target was observed in the photometric mode (“on-off”) of MAMBO which is based on the chop-nod technique and placing the target on a reference bolometer element (on-target channel). The total observing time was 1.5 hours. Data were reduced using MOPSIC, and the current version of MOPSI (Zylka 1998, The MOPSI Cookbook (Bonn: MPIfR)).
The imaging observations were acquired on 10 March 2010 using the dual-arm Low Resolution Imaging Spectrometer [LRIS; (1, 2)] on the 10-m Keck I telescope. Each target field received simultaneous s integrations with the -filter and s integrations with the -filter using the blue and red arms of LRIS. A dither pattern was employed to generate on-sky flat field frames when incorporating all five fields. In addition, s integrations were acquired in the - and -filters for photometric calibration of bright stars in each field. The seeing FWHM for the science exposures was . The data were reduced using IDL routines and combined and analysed using standard IRAF tasks.
Observations of two H-ATLAS/SDP sources, ID81 and ID130, were obtained at 880m using the only current high resolution submillimeter facility in the world, the SMA. The SMA is an interferometer located near the summit of Mauna Kea, Hawaiim and consists of eight 6m diameter radio telescopes. The two H-ATLAS sources were observed in Director’s Discretionary Time (from February to May 2010) in three separate configurations, with baselines spanning a spatial range from 6 to 509 meters, over a total of 4 observing periods (Table S3).
Target observations from each period were interspersed with observations of calibration sources, quasars J0909+013 and J0825+031 (phase) and J0730-116 and J0854+201 (amplitude). The phase calibration targets were typically observed every 7 to 15 minutes, depending upon configuration (a faster cycle was used for the larger configurations).
Calibration of the complex visibility data was performed within the SMA’s MIR package, a suite of IDL-based routines designed for use with SMA data. The initial opacity correction was obtained through application of system temperature to the raw visibility data, a standard practice. Further complex gain corrections, to remove both atmospheric and instrumental amplitude and phase variations, were measured using the calibration quasars, which appear as point sources to the interferometer. Calibration of the absolute flux density scale was performed using measurements of Titan, whose continuum and line structure is known to within about 5% at submillimeter/millimeter wavelengths.
The resulting calibrated visibility data for each source were combined and imaged within the NRAO Astronomical Image Processing System (AIPS). Photometry obtained for the SMA images along with those from PACS, SPIRE, MAMBO are given in table S1.
Plateau de Bure Interferometer observations
The H-ATLAS/SDP sources ID81 and ID130 were observed in the CO J=3-2 and CO J=5-4 lines with the IRAM Plateau de Bure Interferometer (3). Both sources were observed in excellent atmospheric conditions and with the full sensitivity of the six-element array. The observing frequencies were based on the redshifts provided by the CSO/Z-spec spectrometer. The receiver bandwidth was adjusted for maximum sensitivity and the observing frequencies centered in the 1 GHz baseband of the narrowband correlator. Observations of ID81 were made on March 22, 2010 for an effective integration time of 22 min and 14 min, respectively, for the CO J=3-2 and J=5-4 lines. The RF calibration was measured on 3C84, and amplitude and phase calibrations were made on 0823+033. The J=3-2 and J=5-4 transitions in ID130 were observed on March 26 and April 16, 2010, respectively, for an effective integration time of 74 min and 32 min. The RF calibration was measured on 3C273, and amplitude and phase calibrations were made on 0906+015. The absolute flux calibration scales for ID81 and ID130 were established using as primary calibrator MWC349. Data reduction and calibration were made using the GILDAS software package in the standard antenna based mode.
Optical spectroscopic observations
Optical spectroscopic observations of ID11 and ID17 were made using the ISIS double-arm spectrograph on the 4.2-m William Herschel Telescope (WHT). The R158B and R158R gratings were used to provide wavelength coverage across the entire optical spectrum, split by a dichroic at 5300 Å. Four 900-second exposures were taken of each source in a standard ‘ABBA’ pattern, nodding the telescope along the slit by 10 arcseconds between the first and second exposures, and back to the original position between the third and fourth integrations. This allowed initial sky subtraction to be performed by simply subtracting the ’A’ frames from the ’B’ frames. Additional sky subtraction was performed by subtracting the median value of each row, and then the positive and negative beams were aligned and coadded. Wavelength calibration was performed using observations of arc lamps taken with the same set-up. A one-dimensional spectrum was then optimally extracted. The spectra were taken through thin cloud and therefore no attempt has been made to flux-calibrate them. There was very little signal in the blue arms and so only the red-arm spectra are presented here. The redshifts of the two sources were determined by cross-correlation with template spectra. All reduction steps were undertaken using the IRAF package. The resulting spectra are shown in Fig. S3. The spectrum of ID11 reveals absorption lines associated to singly ionised calcium Ca H+K (rest-frame wavelengths: 3968.5 for H-line and 3933.7 for K-line) and the 4000 break feature (rest-frame wavelength 4000) at , while the spectrum of ID17 shows the emission from oxygen doublet [OII]3727 (rest-frame wavelengths 3726-3729) and the 4000 break feature at . In both spectra the absorption feature observed at is due to the Earth’s atmosphere.
A 30-minute exposure of ID130 was taken on May 15, 2010, with the Apache Point Observatory’s 3.5-meter telescope and the DIS [Dual Imaging Spectrograph, (4)] long-slit spectrograph through medium clouds at an average airmass of 1.5. The spectrum was processed by subtracting the detector bias, dividing by a flat-field frame to correct for variable pixel response, performing distortion correction to align the spectrum in the wavelength and spatial directions, subtracting the sky flux determined from parts of the slit containing no sources, and applying a wavelength calibration by reference to emission lines from a Helium-Neon-Argon calibration lamp. Two emission lines in the spectrum (Fig. S4) were identified as [O II]3727 and [Ne III]3869 (rest-frame wavelengths 3869) from the ratio of their observed wavelengths. From the ratio of their observed to emitted wavelengths the redshift of the galaxy was determined to be 0.22010.002.
Modelling with GALFIT
GALFIT (5) is a publicly available two-dimensional non-linear fitting algorithm, which allows galaxy images to be modelled with one or multiple analytical light profiles. Each profile is constrained by a function and a set of parameters. GALFIT convolves the profiles with a user supplied point spread function, in this case empirical point spread functions constructed using nearby stars, and then performs a least-squares minimisation. No hard or soft constraints were applied to the fitting parameters to avoid any prior on the galaxy morphological type. For ID9 and ID11 single Sersic profiles resulted in a reduced close to 1.0 (see table S4 for the best fit parameters). ID17 was fitted with two Sersic component, assuming two lensing galaxies. The resulting Sersic indices were both less than 1 (see table S4). For ID81 and ID130 two components were necessary to achieve a satisfactory fit, with a clean residual. The best fits were obtained using a combination of a compact elliptical Sersic core plus an exponential disk. No detectable background structure was revealed after subtracting the models, which shows the background galaxy is below the optical detection limit. Postage stamp images of ID9, ID11, ID17, ID81 and ID130 are shown in FigsS5 and S6, together with the corresponding best-fit models and residuals, while Fig.S7 shows the individual GALFIT components for ID81 and ID130.
To derive photometric upper limits, we performed random aperture photometry on the i- and g-band Keck maps, using a 1.5 arcsecond radius. This radius was chosen to correspond with the structure visible in the SMA images for ID81 and ID130, which extends to regions with radii of approximately 11.5 arcseconds. The resulting flux distributions were fitted with Gaussians and the 3 upper limits are presented in Table S4.
Mass estimate from lensing
The Einstein radius of a strong galaxy-galaxy gravitational lens system can be measured from the configuration of multiple lensed images by averaging the distances of the images from the center of the lensing galaxy. For two of the H-ATLAS/SDP lens candidates, ID81 and ID130, the positions of the lensed images are constrained by high-resolution SMA follow-up imaging. The lensed images of the background sources appear as peaks in the SMA signal-to-noise ratio map. Here we have selected those peaks with signal-to-noise ratio above eight, which provided positions for four images in ID81 and two images in ID130. The error on the Einstein radius is estimated by taking into account the uncertainties on the position of the individual peaks. For a point source the rms error on its position is /SNR (assuming no systematic astrometry errors and uncorrelated Gaussian noise), where is the Gaussian rms width of the instrument beam (=FWHM/), while SNR is the signal-to-noise ratio at the source position (6, 7). The SMA synthesised beam (derived by combining observations in VEX, COM and SUB configurations) has size 0.810.73 for ID81 and 0.780.72 for ID130. Therefore, in estimating the relative positional uncertainty of the peaks, we have assumed FWHM=0.75 and FWHM=0.77 for ID81 and ID130, respectively. The absolute positional uncertainty of the SMA images is estimated by referencing the data to nearby point-like sources (quasars) of known positions and is below 10 milli-arcseconds.
Once the Einstein radius is known, the mass within the Einstein ring can be easily derived assuming a Singular Isothermal Sphere (SIS) model (although the result is only little dependent on the model used) which is characterized by a projected surface density that falls off as , where is the angular distance from the center of the mass distribution (8),
and is the critical surface density:
In the equation above, is the speed of light, is the gravitational constant, and are the angular diameter distances to the lens and the source, respectively, while is the angular diameter distance between the lens and the source. The error on the mass is obtained by propagating the errors on the Einstein radius and on the spectroscopic/photometric redshifts used to derived the angular diameter distances. The estimated values of and are listed in Table2.
Gravitational lensing modeling
A detailed analysis of the lensed structure revealed by the SMA images is beyond the scope of this paper and is deferred to a forthcoming publication. However, in order to prove that such a structure is consistent with a lensing event, we have used the publicly available LENSMODEL software (9) to reproduce the positions of the peaks in the SMA maps. We have assumed a Singular Isothermal Ellipsoid (SIE) model (8) for the mass distribution of the lens galaxy. The SIE model consists of concentric and aligned elliptical isodensity contours with axis ratio . The circular limit is the SIS model and corresponds to . The results are shown in Fig.S9. We have further assumed that the centroid of the mass model coincides with that of the light distribution of the lensing galaxy. The best-fit model for ID130 has ellipticity and position angle (measured East of North) of deg, consistent with the results found for the optical light-distribution that is dominated by the more compact Sersic profile (TableS4 and Fig.S8). For ID81, the mass distribution has ellipticity and position angle deg, which is not consistent with that measured for the luminous component (TableS4 and Fig.S8). Besides, the position of the peaks is not well reproduced by the model. This may hint at the effect of an external shear (which we did not include) due to a nearby cluster (photometrically detected 3.6 arcminutes away), in the direction indicated by the arrow in Fig.S9.
We have used the best-fit lens models to approximatively quantify the magnification experienced by a background source described by a Gaussian profile with a Full Width at Half Maximum (FWHM) in the range 0.1-0.3. This extension is consistent with the physical size of the submillimeter galaxy studied by (10). The inferred magnification is 18-31 for ID81 and 5-7 for ID130. An example of lensed image, after convolution with the SMA point spread function, for the case FWHM=0.2 is shown in Fig.S9.
|H-ATLAS ID||SDP ID|
|GALEX FUV (Jy)||-||-||-||0.230.18||-|
|GALEX NUV (Jy)||-||-||-||1.91.1||-|
|SDSS u (Jy)||0.240.23||0.570.59||3.31.6||3.92.0||1.71.7|
|SDSS g (Jy)||1.790.43||1.010.45||3.96.4||24.91.1||19.410.72|
|SDSS r (Jy)||5.810.70||3.940.65||7.71.0||114.82.1||66.11.2|
|SDSS i (Jy)||14.91.1||11.31.0||15.31.5||197.73.6||108.62.0|
|SDSS z (Jy)||27.03.7||21.54.2||11.86.0||278.03.6||143.26.6|
|UKIDSS Y (Jy)||-||-||27.79.5(6.6)||321.33.2(6.3)||-|
|UKIDSS J (Jy)||-||102.49.8(16)||5617(12)||36711(9.2)||-|
|UKIDSS H (Jy)||7315(5.0)||23717(14)||10719(8.2)||508.15.3(8.5)||-|
|UKIDSS K (Jy)||13224(6.5)||-||10823(9.7)||573.76.2(14)||-|
References and Notes
J. B. Oke, et al., Publ. Astron. Soc. Pac. 107, 375 (1995).
J. K. McCarthy, et al., Society of Photo-Optical Instrumentation Engineers (SPIE) Conference, S. D’Odorico, ed. (1998), vol. 3355, pp. 81-92
S. Guilloteau, et al., Astron. & Astrophys. 262, 624 (1992)
C. Y. Peng, L. C. Ho, C. D. Impey, H. Rix, Astron. J. 124, 266 (2002)
J. J. Condon, Publ. Astron. Soc. Pac. 109, 166 (1997)
R. J. Ivison, et al., Mon. Not. R. Astron. Soc. 380, 199 (2007)
R. Kormann, P. Schneider, M. Bartelmann, Astron. & Astrophys. 284, 285 (1994)
C. R. Keeton, available at http://xxx.lanl.gov/abs/astro-ph/0102340 (2001)
A. M. Swinbank, et al., Nature 464, 733 (2010)
D. Hill, et al., available at http://xxx.lanl.gov/abs/astro-ph/1009.0615
- Total bandwidth coverage is LO-8 to LO-4 (LSB) and LO+4 to LO+8 (USB) for a total of 8 GHz. The small difference in LO Frequency between compact configuration observations and the subcompact and very extended observations is not important in this context.
- ExpDisk = exponential disk profile
- Radius for Sersic and disk scale length for ExpDisk
- Sersic index
- Angle measured east of north
- When multiple images of the same background source are formed, the event is known as strong gravitational lensing .
- P. J. Marshall, et al., Astrophys. J. 671, 1196 (2007).
- D. P. Stark, et al., Nature 455, 775 (2008).
- A. M. Swinbank, et al., Nature 464, 733 (2010).
- I. Smail, R. J. Ivison, A. W. Blain, Astrophys. J. L. 490, L5+ (1997).
- A. W. Blain, J. Kneib, R. J. Ivison, I. Smail, Astrophys. J. L. 512, L87 (1999).
- M. Rowan-Robinson, et al., Nature 351, 719 (1991).
- I. Smail, R. J. Ivison, A. W. Blain, J. Kneib, Mon. Not. R. Astron. Soc. 331, 495 (2002).
- J. D. Vieira, et al., Astrophys. J. 719, 763 (2010).
- G. L. Pilbratt, et al., Astron. & Astrophys. 518, L1+ (2010).
- A. W. Blain, Mon. Not. R. Astron. Soc. 283, 1340 (1996).
- F. Perrotta, C. Baccigalupi, M. Bartelmann, G. De Zotti, G. L. Granato, Mon. Not. R. Astron. Soc. 329, 445 (2002).
- F. Perrotta, et al., Mon. Not. R. Astron. Soc. 338, 623 (2003).
- M. Negrello, et al., Mon. Not. R. Astron. Soc. 377, 1557 (2007).
- K. Coppin, et al., Mon. Not. R. Astron. Soc. 372, 1621 (2006).
- S. C. Chapman, A. W. Blain, I. Smail, R. J. Ivison, Astrophys. J. 622, 772 (2005).
- S. Serjeant, D. Harrison, Mon. Not. R. Astron. Soc. 356, 192 (2005).
- G. de Zotti, et al., Astron. & Astrophys. 431, 893 (2005).
- L. V. E. Koopmans, T. Treu, A. S. Bolton, S. Burles, L. A. Moustakas, Astrophys. J. 649, 599 (2006).
- I. W. A. Browne, The Class Collaboration, Gravitational Lensing: Recent Progress and Future Go, T. G. Brainerd & C. S. Kochanek, ed. (2001), vol. 237 of Astronomical Society of the Pacific Conference Series, pp. 15–+.
- M. Oguri, et al., Astron. J. 132, 999 (2006).
- C. Faure, et al., Astrophys. J. S. 176, 19 (2008).
- A. Weiß, et al., Astrophys. J. 707, 1201 (2009).
- M. J. Devlin, et al., Nature 458, 737 (2009).
- S. Eales, et al., Publ. Astron. Soc. Pac. 122, 499 (2010).
- M. J. Griffin, et al., Astron. & Astrophys. 518, L3+ (2010).
- E. Pascale, submitted; preprint available at http://xxx.lanl.gov/abs/astro-ph/1010.5782 .
- A. Poglitsch, et al., Astron. & Astrophys. 518, L2+ (2010).
- E. Ibar, submitted; preprint available at http://xxx.lanl.gov/abs/astro-ph/1009.0262 .
- S. Driver, submitted; preprint available at http://xxx.lanl.gov/abs/astro-ph/1009.0614 .
- E. E. Rigby, submitted; preprint available at http://xxx.lanl.gov/abs/astro-ph/1010.5787 .
- D. L. Clements, et al., Astron. & Astrophys. 518, L8+ (2010).
- M. Baes, et al., Astron. & Astrophys. 518, L39+ (2010).
- J. González-Nuevo, et al., Astron. & Astrophys. 518, L38+ (2010).
- Supporting online material (SOM) .
- D. J. B. Smith, et al., submitted; preprint available at http://xxx.lanl.gov/abs/astro-ph/1007.5260 .
- D. H. Hughes, et al., Mon. Not. R. Astron. Soc. 335, 871 (2002).
- I. Aretxaga, et al., Mon. Not. R. Astron. Soc. 342, 759 (2003).
- A. Weiß, et al., Astrophys. J. L. 705, L45 (2009).
- E. Daddi, et al., Astrophys. J. L. 695, L176 (2009).
- B. J. Naylor, et al., Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, T. G. Phillips & J. Zmuidzinas, ed. (2003), vol. 4855 of Presented at the Society of Photo-Optical Instrumentation Engineers (SPIE) Conference, pp. 239–248.
- C. M. Bradford, et al., Astrophys. J. 705, 112 (2009).
- R. Lupu, submitted; preprint available at http://xxx.lanl.gov/abs/astro-ph/1009.5983 .
- A. I. Harris, et al., From Z-Machines to ALMA: (Sub)Millimeter Spectroscopy of Galaxies, A. J. Baker, J. Glenn, A. I. Harris, J. G. Mangum, & M. S. Yun , ed. (2007), vol. 375 of Astronomical Society of the Pacific Conference Series, pp. 82–+.
- The Zpectrometer frequency coverage allows it to search for CO J=10 transition over the redshift range 2.2 to 3.5 simultaneously .
- D. Frayer, submitted; preprint available at http://xxx.lanl.gov/abs/astro-ph/1009.2194 .
- Only g and i-band photometry can be obtained for each of the two galaxies detected in ID17. For this reason, we did not attempt any fit to the optical/NIR SED of this H-ATLAS source .
- E. da Cunha, S. Charlot, D. Elbaz, Mon. Not. R. Astron. Soc. 388, 1595 (2008).
- E. da Cunha, et al., in press; preprint available at http://xxx.lanl.gov/abs/astro-ph/1008.2000 .
- The visual extinction parameter, , is the difference between the observed V-band magnitude and the intrinsic (dust-free) V-band magnitude of the galaxy .
- S. M. Faber, R. E. Jackson, Astrophys. J. 204, 668 (1976).
- C. J. Willott, S. Rawlings, M. J. Jarvis, K. M. Blundell, Mon. Not. R. Astron. Soc. 339, 173 (2003).
- J. Huang, L. L. Cowie, G. A. Luppino, Astrophys. J. 496, 31 (1998).
- D. H. Gudehus, Astrophys. J. 382, 1 (1991).
- M. W. Auger, et al., Astrophys. J. 705, 1099 (2009).
- Herschel is an ESA space observatory with science instruments provided by European-led Principal Investigator consortia and with important participation from NASA. US participants in H-ATLAS acknowledge support from NASA through a contract from JPL. This work was supported by the Science and Technology Facilities Council (grants PP/D002400/1 and ST/G002533/1) and studentship SF/F005288/1. We thank Agenzia Spaziale Italiana for funding through contract No. I/016/07/0 COFIS and ASI/Istituto Nazionale di Astrofisica agreement I/072/09/0 for the Planck Low Frequency Instrument (LFI) Activity of Phase E2. Research supported in part by Consejo Nacional de Ciencia y Tecnologia (CONACyT) grants 39953-F and 39548-F. The W.M. Keck Observatory is operated as a scientific partnership among the California Institute of Technology, the University of California and NASA. The Observatory was made possible by the generous financial support of the W.M. Keck Foundation. The Submillimeter Array is a joint project between the Smithsonian Astrophysical Observatory and the Academia Sinica Institute of Astronomy and Astrophysics and is funded by the Smithsonian Institution and the Academia Sinica. IRAM is supported by Institut National des Sciences de l’Univers (INSU)/CNRS (France), Max Planck Society (MPG) (Germany) and Instituto Geografico Nacional (Spain). Z-spec was supported by NSF grant AST-0807990 to J. A. and by the CSO NSF Cooperative Agreement AST-0838261. Support was provided to J. K. by an NSF Graduate Research Fellowship. Z-spec was constructed under NASA SARA grants NAGS-11911 and NAGS-12788 and an NSF Career grant (AST-0239270) and a Research Corporation Award (RI0928) to J. G., in collaboration with the JPL, California Institute of Technology, under a contract with NASA. Construction of and observations with the Zpectrometer have been supported by NSF grants AST-0503946 and AST-0708653. The NRAO is a facility of the NSF operated under cooperative agreement by Associated Universities. The optical spectroscopic redshift of ID130 was derived from observations obtained with the Apache Point Observatory 3.5-meter telescope, which is owned and operated by the Astrophysical Research Consortium. The optical spectroscopic redshifts of ID9 and ID11 were obtained with the William Herschel Telescope which is operated on the island of La Palma by the Isaac Newton Group in the Spanish Observatorio del Roque de los Muchachos of the Instituto de Astrofísica de Canarias. For the use of Keck, SMA and CSO, the authors wish to recognize and acknowledge the very important cultural role and reverence that the summit of Mauna Kea has always had within the indigenous Hawaiian community. We are most fortunate to have the opportunity to conduct observations from this mountain .