ALMACAL IV: A catalogue of ALMA calibrator continuum observations
We present a catalogue of ALMA flux density measurements of 754 calibrators observed between August 2012 and September 2017, for a total of 16,263 observations in different bands and epochs. The flux densities were measured reprocessing the ALMA images generated in the framework of the ALMACAL project, with a new code developed by the Italian node of the European ALMA Regional Centre. A search in the online databases yielded redshift measurements for 589 sources (78 per cent of the total). Almost all sources are flat-spectrum, based on their low-frequency spectral index, and have properties consistent with being blazars of different types. To illustrate the properties of the sample we show the redshift and flux density distributions as well as the distributions of the number of observations of individual sources and of time spans in the source frame for sources observed in bands 3 (84116 GHz) and 6 (211275 GHz). As examples of the scientific investigations allowed by the catalogue we briefly discuss the variability properties of our sources in ALMA bands 3 and 6 and the frequency spectra between the effective frequencies of these bands. We find that the median variability index steadily increases with the source-frame time lag increasing from 100 to 800 days, and that the frequency spectra of BL Lacs are significantly flatter than those of flat-spectrum radio quasars. We also show the global spectral energy distributions of our sources over 17 orders of magnitude in frequency.
keywords:galaxies: photometry – galaxies: active – galaxies: abundances – submillimetre: galaxies
The Atacama Large Millimeter/submillimeter Array (ALMA) calibrators comprise many hundreds of bright, compact radio sources, distributed over about 85 per cent of the sky.
Every ALMA science project includes observations of calibrator sources (mostly bright quasars in the mm and sub-mm regime) to set the flux density scale, to measure the bandpass response, and to calibrate amplitude and phase of the visibilities of the science targets (Fomalont2014).
Such observations represent a significant fraction (generally 30 per cent) of each execution block (EB). If the calibrator is a phase calibrator, it is observed many times during the same EB. If it is a bandpass or an amplitude calibrator, it is typically observed once per EB. Therefore each calibrator can be observed several times, on different dates, in different ALMA bands and array configurations, for one or multiple science projects.
The fields around ALMA calibrators of projects stored in the ALMA Science Archive have been exploited to carry out a novel, wide and deep (sub-)millimetre survey, ALMACAL (Oteo2016), and to investigate detected sources of special interest (Oteo2017; Klitsch2017). The ALMACAL survey, in fact, takes advantage of the high sensitivity reached in the fields of ALMA calibrator observations to blindly extract a multi-band, multi-epoch survey of dusty star forming galaxies. Together with this primary goal, the same observations offer a unique opportunity to investigate spectral behaviour and variability of a large sample of bright extragalactic sources, the calibrators themselves, mostly AGNs across the whole (sub-)millimetric band.
In this paper we present a catalogue of observations of the ALMA calibrators collected, so far, for the ALMACAL project purposes. Their multi-epoch, multi-frequency measurements over a poorly explored spectral region constitute a rich database, well-suited for a variety of scientific investigations, some of which will be more extensively detailed in future papers of our collaboration.
This paper is structured as follows. In Section 2, the catalogue is introduced. An account of the source classification is given in Section 3, where we also compare the frequency spectra of BL Lacs and flat-spectrum radio quasars (FSRQs). In Section 4, we describe the main properties of the catalogue and, as an example of its scientific exploitation, we briefly discuss the variability properties of our sources in ALMA bands 3 and 6. In Section 5, we present the global spectral energy distributions (SEDs) of sources, built collecting data from online databases. Finally, Section 6 contains a short summary of the paper.
|ALMA name||Class.||z||RA [deg]||DEC [deg]||Flux density [Jy]||Error [Jy]||band||[GHz]||Date of obs.|
|Notes. Classification: 1=Flat-spectrum radio quasar (FSRQ); 2=BL Lac; 3=BL Lac-galaxy dominated; 4=Blazar uncertain type;|
|5=BL Lac candidate; 6=Steep spectrum; 7=Uncertain.|
|The uncertainty is given by summing in quadrature the r.m.s. and a typical ALMA calibration error equal to 5 per cent of|
|the flux (see text).|
|Observing time in the format [YYYY/MM/DD/hh:mm:ss], UTC time.|
2 The sample
The catalogue consists of continuum measurements of calibrators, obtained during the majority of the ALMA science observations between August 2012 and September 2017. In total, we collected 16,263 observations111From the initial sample, we removed 1.7 per cent of the images that showed anomalies. of 754 calibrators. Being a collection of data from a heterogeneous sample of science projects, our observations vary within a wide spectrum of different frequency setups, array configurations and integration times.
The details of calibration and imaging for the ALMACAL data are described by Oteo2016. Here we summarise a few pieces of information, useful for a comprehensive description of the presented catalogue. For our purposes we considered all the ALMA projects in the epochs 2012-2017 and, in them, the extragalactic calibrators at any observing band. The full data deliveries available in the ALMA archive were retrieved for datasets for which the proprietary period had expired, while only the calibrator data was considered for the remaining projects, after an official request through an ALMA Helpdesk ticket. Calibration scripts produced during the ALMA Quality Assessment procedure and distributed through the archive were run to generate the calibration tables that were applied to all the calibrators (in some cases differently with respect to what usually done for the archived data, for which, expecially in the first observing cycles, tables were applied only to science targets and phase calibrators). Data were self-calibrated taking advantage of the presence of the calibrator in the phase center and images are produced with the calibrator present and subtracted (in the visibility domain). The latter are used in the ALMACAL collaboration to investigate the background looking for dusty galaxies. The former are used in the present paper to investigate the calibrator population properties.
For the calibrators, the flux densities were uniformly measured from the ALMA images222We derived the flux densities through an image analysis, instead of simply using model fit values, because the former approach provides robust measurements for both resolved and non-resolved sources, while model fit flux densities are reliable for non-resolved observations only using a new code developed by the Italian node of the European ALMA Regional Centre (ARC). This software is part of a suite of tools aimed at easing the ALMA Science Archive mining: the ALMA Keyword Filler tool package (AKF; Liuzzo18) and the Keywords of Astronomical FITS-images Explorer (KAFE; Burkutean18). The AKF codes are particularly useful to compare image products or to identify the images to be selected for several scientific purposes. KAFE is a web-based FITS image post-processing analysis tool. It exploits AKF and complements selected FITS files with metadata based on a uniform image analysis approach while also offering advanced image diagnostic plots. KAFE’s applicability to multi-instrument images in the radio to sub-mm wavelength domain makes it ideal for data sample studies requiring uniform data diagnostic criteria.
After the estimation of the r.m.s. () in an image, the code masks the pixels with a flux density below and obtains the source flux density by integrating over the remaining pixels. This is enough to cope both with isolated point-like and extended sources, definition which strongly depends on the observing strategy and phase decoherence and might vary for our targets from one observation to the other.
The number of observations in the different ALMA bands are: 5100 in band 3 (84116 GHz), 639 in band 4 (125163 GHz), 6319 in band 6 (211275 GHz), 3584 in band 7 (275373 GHz), 393 in band 8 (385500 GHz), 220 in band 9 (602720 GHz) and 8 in band 10 (787950 GHz).
The ALMA measurements of the 754 calibrators are included as supplementary material in the electronic version of the paper and on the website of the Italian ARC (http://arc.ia2.inaf.it). The catalogue gives the ALMA name, the source classification, its redshift (if available), the equatorial coordinates (J2000), the flux density measured in each observation with its error, the effective observing frequency, and the date and UTC time of the observation. The error is essentially given by the uncertainty in the flux density calibration (errors due to instrumental noise are typically smaller by more than two orders of magnitudes); we adopt a calibration uncertainty of 5 per cent (E. Fomalont, private communication)333The debate about the precise value of the calibration uncertainty is still open in the ALMA community. Our results about the differences in flux density of the different calibrators for short time spans (see Sect. 4) support the adopted 5 per cent level.. An example of the content of the catalogue is given in Table 1. The coordinates are the average between the positions measured in the different ALMA observations.
We have recovered the redshifts of 589 sources (78 per cent of the total), using the Astroquery444https://astroquery.readthedocs.io/en/latest/ affiliated package of astropy555http://www.astropy.org/ on the NASA/IPAC Extragalactic database666https://ned.ipac.caltech.edu/ (NED), VizieR777http://vizier.u-strasbg.fr/viz-bin/VizieR and SIMBAD888http://simbad.u-strasbg.fr/simbad/ databases. Redshifts for 256 calibrators were provided by Mahony2011. Whenever multiple redshifts of the same source were found, we give the median value.
In Fig. 1, we show the Mollweide projection of the positions of the ALMA calibrators, obtained through KAFE.
3 Source classification
Most sources of our catalogue (489, i.e. 67 per cent) are included in the 5th edition of the Roma Multi-frequency Catalogue of Blazars999http://www.ssdc.asi.it/bzcat/ (BZCAT; Massaro2009). BZCAT sources are divided into 5 sub-classes: FSRQs, BL Lacs, BL Lacs-galaxy dominated, Blazars of uncertain type, BL Lac candidates.
We split the remaining 265 sources into the two classical sub-populations of steep-spectrum and flat-spectrum sources. As usual, such classification is based on the low-frequency (between and GHz) spectral index, , adopting () as the boundary value. The flat-spectrum population is essentially made by blazars.
We computed using the 1.4 GHz flux densities from the NRAO VLA Sky Survey (NVSS; Condon1998) complemented with those at 843 MHz from the Sydney University Molonglo Sky Survey (SUMSS; Mauch2003), combined with those at 4.85 GHz from the Green Bank 6 cm (GB6; Gregory1996) or from the Parkes-MIT-NRAO (PMN; GriffithWright1993) survey catalogues.
The low frequency spectral index could be computed for all but 13 sources (out of 265), that were classified as “uncertain”. Sources with were classified as steep-spectrum, provided they did not show clear variability or -ray emission. Only 10 sources satisfy the criteria for a steep-spectrum classification.
The overwhelming majority, 731 sources, i.e. 97 per cent of the sample, are classified as blazars (since they belong to the BZCAT catalogue or they fulfill the criteria presented above). This includes also those with but with statistically significant variability and/or -ray emission (31 sources). We classify as “Blazar uncertain type” our blazars without a BZCAT classification.
The classical physical models of blazars predict a steepening of their radio spectra at millimeter wavelengths (Kellermann1966; Blandford1979). Statistical evidence of such steepening has been reported by several authors (GonzalezNuevo2008; PlanckCollaborationXIII2011; PlanckCollaborationXV2011; PlanckCollaborationXLV2016). The ALMA data allow us to check this prediction on a much larger sample than was possible before.
Tucci2011 went one step further. Their most successful physical evolutionary models of radio sources entail different distributions of break frequencies (the frequencies where the spectra steepen), for BL Lacs and FSRQs. They argue that BL Lacs have substantially higher break frequencies, implying that their synchrotron emission comes from more compact regions. Their best model, C2Ex, that successfully fits number counts and spectral index distributions of extragalactic radio sources over the 5–220 GHz frequency range, predicts, for bright blazars (Jy, like sources in our sample), that the break frequencies of most FSRQs are well below 100 GHz while those of most BL Lacs are well above this frequency (cf. their Fig. 7). The ALMA data are well suited to test this prediction.
The left panel of Fig. 3 shows the distribution of high-frequency spectral indices (, from the effective frequencies of ALMA band 3 [84116 GHz] to those of band 6 [211275 GHz]) versus for FSRQs and BL Lacs. The right panel shows the distributions of such indices.
Most low-frequency spectral indices are in the range from to 0.8, while most of the high-frequency ones range from to 0. Within these ranges there is no correlation between the high- and low-frequency spectral indices. The median spectral indices substantially steepen from low to high frequencies. For FSRQs we have (with first and third quartile values of about 0.10 and 0.29) and (first and third quartile values of about 0.85 and 0.51). For BL Lacs (first and third quartile values of about 0.18 and 0.22) and (first and third quartile values of about 0.61 and 0.33). The global (FSRQ + BL Lac) median high-frequency spectral index (with first and third quartile values of about 0.80 and 0.45) is in good agreement with those found by Massardi2016 for the Planck–ATCA Co-eval Observations (PACO) bright sample: (with first and third quartile values of and ); (with first and third quartile values of and ).
There is thus evidence of a flatter median of BL Lacs compared to FSRQs. The statistical significance of the difference was estimated using the two-sample Kolmogorov-Smirnov (KS) test, i.e. computing
where is the KS statistics, that is the largest discrepancy between the cumulative distributions of high-frequency spectral indices of the two source populations, FSRQs and BL Lacs, comprising and sources, respectively.
We find corresponding to a per cent probability that the two populations are drawn from the same parent distribution. A simpler, although less rigorous, illustration of the significance of the difference can be obtained considering that the ratio of the numbers of FSRQs in the bins and ) is (the FSRQs in the two bins are ; see the right panel of Fig. 3). If the BL Lacs were extracted from the same parent population we would expect a similar ratio between the two bins, i.e. the expected number of BL Lacs in the first bin would be and in the second bin would be . Based on the Poisson statistics, the probability of getting in the second bin 20 objects when 10.4 are expected is per cent, close to the result of the KS test.
The statistically significant difference between the distributions of for the two populations might be consistent with higher break frequencies for BL Lacs compared to FSRQs, as suggested by the Tucci2011 model. However PlanckCollaborationXLV2016 did not find significant differences in the break frequencies of the two populations for their complete flux-density-limited sample of 104 extragalactic radio sources detected by the Planck satellite, but reported average spectral indices above the break frequency significantly steeper for FSRQs than for BL Lacs.
4 Properties of the sample
The flux density distributions of sources in the two most frequented ALMA bands (bands 3 and 6) are shown in Fig. 4. They extend from 1 mJy to 15 Jy, with a peak at 0.2 Jy. For comparison, the minimum flux densities of sources in the “extragalactic zone” () listed in the Second Planck Catalogue of Compact Sources (PCCS2; PCCS2) are 232 mJy at 100 GHz and 127 mJy at 217 GHz. Therefore, the ALMA observations reach much fainter flux densities than the Planck ones, but there is a large overlap between the two sets of observations.
As mentioned in Sect. 2 we have recovered redshift measurements for 78 per cent of our sample (589 sources). The redshift distributions of sources detected in bands 3 and 6 are shown in Fig. 5. Both distributions peak at and have tails extending up to .
Figure 6 shows the distribution of the number of observations of a given source, again in bands 3 and 6, for sources with 2 observations. The distributions peak at the lowest bin (2-8 measurements) but there is a significant number of sources with tens of measurements, up to ; sources with 20 observations are 46 in band 3 and 52 in band 6.
The time span distribution of measurements in the source frame (, where is the time between the first and the last observation) is shown in Fig. 7. We excluded all the sources lacking redshift measurements. Although for most sources the time span is relatively short (less than a few hundreds of days), for some sources observations cover a few years (in the source frame); sources with years are 20 in band 3 and 28 in band 6.
The light curves in bands 3 and 6 of three of the most frequently observed sources are shown in Fig. 8. Monitoring of blazars is important to understand which mechanisms drive their violent variability and what is the duty cycle of their activity. The interest on multi-frequency blazar monitoring has gained momentum since the launch of the Fermi Gamma-ray Space Telescope. In fact the overwhelming majority of detected extragalactic -ray sources are blazars and Fermi has gathered spectacular -ray light curves of hundreds of them (Abdo2010; Abdollahi2017). The poorly explored few-mm to sub-mm spectral region covered by ALMA observations is important in this context since flux densities in this region appear to be particularly well correlated with those at 1 GeV (Fuhrmann2016).
An obvious use of multiple observations is the calculation of the variability index (), defined as (Sadler2006):
where and are the flux density measurements of a source measured in a given band and the associated uncertainties, is the number of measurements, and is the mean flux density.
Obviously, the variability index can be reliably measured only if the amplitude of flux density variations is substantially larger than the 5 per cent calibration uncertainty, although Eq. (2) gives values of the variability index 5 per cent. Reliable variability indices are measured for 31 out of 41, 30/33, 29/33 and 26/26 (band 3) and 25/30, 36/39, 35/39 and 37/37 (band 6) sources for time spans of 100, 200, 400 and 800 days (within 30 per cent, in the source frame), respectively. All sources with measurements on these timescales are included.
In Table 2 the median, first quartile and third quartile values of the VI for the different source frame timescales and in the two different bands are listed.
Applying the two-sample KS test to the VI distributions of the 100 and 800 days of timescales, we find that the probability that the two sub-samples are drawn from the same parent distribution is extremely low (0.1 per cent in both bands, with D0.588 in band 3 and D0.732 in band 6). This is a direct consequence of the fact that the characteristic timescale of blazar variability in blazar sources is 3 years (see e.g. Nieppola2009).
Measured flux densities of the same source for short time spans are expected to be only weakly affected by variability. Differences among such measurements are therefore an estimator of systematic errors that afflict our observations, and primarily of the calibration error. The median absolute values of differences among measurements in bands 3 and 6 done within 30 days in the source frame are of per cent for band 3 and of per cent for band 6, consistent with the adopted calibration error of 5 per cent.
5 Global spectral energy distributions (SEDs) of sources
As mentioned in Sect. 3, our source classification is based on external data. For each source in our sample, we have collected the photometric data available on the NASA/IPAC Extragalactic Database (NED) using Astroquery with a search radius of 10 arcsec and excluding Galactic sources.
The NED data were complemented by cross-matching our catalogue with: the Australia Telescope 20GHz Survey Catalog (AT20G; Murphy2010), the PCCS2 (PCCS2), the PACO catalogue (Massardi2016), the GaLactic and Extragalactic All-sky MWA survey (GLEAM; HW2017), the CRATES survey (Healey07).
The cross-matching was done using the following search radii: 5 arcsec around the ALMA positions for the AT20G catalog; 16, 13.5, 6.5, 4.85, 3.6, 2.45, 2.45, 2.35 and 2.1 arcmin (i.e. half FWHM) for the 30, 44, 70, 100, 143, 217, 353, 545 and 857 GHz PCCS2 catalogues, respectively; 20 arcsec for the GLEAM catalog; 70 arcsec for the CRATES one. For the PACO catalog we exploited the AT20G identifications by Massardi2016. We considered PCCS2 data only for sources with to avoid wrong identifications (with Galactic sources). In all the cases, the search yielded a unique identification.
We have presented a catalogue of ALMA flux density measurements of calibrators, observed between August 2012 and September 2017, in the framework of the ALMACAL project. The ALMACAL images were reprocessed using a new code developed by the Italian node of the European ALMA Regional Centre. This has yielded 16,263 flux density measurements in different ALMA bands and at different epochs of 754 calibrators. A search in online databases has yielded redshifts for 589 sources (78 per cent of the total).
Most (489, i.e. 67 per cent) of our sources are classified as blazars of various types in the BZCAT catalogue. Almost all of the remaining sources have properties (flat low-frequency radio spectrum, clear variability in different bands, -ray emission) consistent with a blazar classification. In total, 97 per cent of the sources are classified as blazars.
To illustrate the properties of the sample, in view of its exploitation for scientific investigations, we have focussed on the most frequented ALMA bands, i.e. bands 3 and 6. For these bands we have shown the redshift and flux density distributions of catalogued sources, the distribution of the number of observations of individual sources and of time spans in the source frame.
Several sources have tens of measurements in a band, covering several years. As an example of the variety of scientific investigations allowed by the catalogue, we have presented unprecedented band 3 and 6 light curves of three sources and estimates of the variability indices on timescales of 100, 200, 400 and 800 days in the same bands.
Through an analysis of flux density differences for short time spans, in bands 3 and 6, we have found that the systematic errors are consistent with the adopted calibration error of 5 per cent.
We have also found that the ALMA data show highly significant evidence of a difference between the high-frequency (GHz) spectra of FSRQs and BL Lacs: at wavelengths a few mm the average spectra of BL Lacs are flatter than those of FSRQs. This is expected if the synchrotron emission of BL Lacs comes from more compact regions than the emission of FSRQs, as argued, e.g., by Tucci2011.
Finally, by collecting data from online databases, we have reconstructed the SEDs of our sources over 17 orders of magnitude in frequency. Both the catalogue and the SEDs are available to the community.
We thank Ian Smail for his useful comments. We are indebted to Edward Fomalont for enlightening discussions, highly valuable suggestions and for his extensive work, together with Ruediger Kneissl and Antonio Hales, on ALMA calibrator catalogs and calibration error estimation. This paper and the AKF and KAFE development are part of the activities for the ALMA Re-Imaging Study approved in the framework of the 2016 ESO Call for Development Studies for ALMA Upgrade (PI: Massardi). The study acknowledges partial financial support by the Italian Ministero dell’Istruzione, Università e Ricerca through the grant ‘Progetti Premiali 2012 - iALMA’ (CUP C52I13000140001). GDZ acknowledges support from ASI/INAF agreement n. 2014-024-R.1 for the Planck LFI Activity of Phase E2 and from the ASI/Physics Department of the university of Roma–Tor Vergata agreement n. 2016-24-H.0 for study activities of the Italian cosmology community. MN acknowledges support from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 707601. This research has made use of the NASA/IPAC Extragalactic Database (NED), which is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration.