EDP Sciences
Free Access
Issue
A&A
Volume 594, October 2016
Article Number A60
Number of page(s) 12
Section Extragalactic astronomy
DOI https://doi.org/10.1051/0004-6361/201628775
Published online 13 October 2016

© ESO, 2016

1. Introduction

The vast majority of high-energy (HE, 100 MeV <E< 100 GeV) and very high-energy (VHE, E> 0.1 TeV) γ-ray sources are associated with radio loud objects, typically blazars, i.e., flat spectrum radio quasars (FSRQs) or BL Lac type objects (BL Lacs; e.g., Acero et al. 2015; Ackermann et al. 2015). Depending on the position of the peak of the synchrotron component of their spectral energy distribution (SED), blazars are further classified as low-, intermediate-, or high-synchrotron peaked (LSP, ISP, HSP) sources, characterized by a synchrotron peak frequency νpeak (Hz) such that log νpeak< 14, 14 < log νpeak< 15, and log νpeak> 15, respectively (Abdo et al. 2010a).

In the MeV/GeV domain, the Large Area Telescope (LAT) on board the Fermi Gamma-ray telescope (Fermi) has provided a deep, uniform sky survey detecting as many as 3,033 sources in the Fermi third source catalog (3FGL, Acero et al. 2015). The results from Fermi show that both types of blazars are common HE emitters, with FSRQs having softer spectra and being generally more luminous than BL Lacs. Moreover, Fermi clearly revealed the existence of a highly significant correlation between the radio flux density and the γ-ray energy flux (Ackermann et al. 2011). However, observations from Imaging Atmospheric Cherenkov Telescopes (IACTs) preferentially detect BL Lac sources, in particular of the HSP type, and no evidence of a correlation between radio and VHE emission has been reported so far. The different demographics of the VHE population is explained by two main factors: FSRQs have softer spectra and are more distant, which decreases their VHE emission owing to extragalactic background light (EBL) absorption. The lack of a VHE-radio correlation may be due to the IACTs operational mode. The IACTs operate in pointing mode with a limited sky coverage, and they preferentially observe sources in a peculiar state. All of these limitations introduce a strong bias in VHE catalogs, and it is difficult to assess any possible radio-VHE correlation. Nonetheless, it is likely that some physical effect may also be at work.

The first Fermi-LAT catalog of sources above 10 GeV (1FHL, Ackermann et al. 2013) is an ideal resource for addressing the connection between HE and VHE emission. The 1FHL is based on LAT data accumulated during the first three years of the Fermi mission, providing a deep and uniform all-sky survey. It contains 514 sources, of which ~76% are statistically associated with active galactic nuclei (AGN), ~11% are sources of Galactic nature (pulsars, supernova remnants, and pulsar wind nebulae), while ~13% remain unassociated. Various observational campaigns were dedicated to searching for the low-frequency counterpart of the unassociated γ-ray sources (UGS) detected by Fermi (Nori et al. 2014; Massaro et al. 2013b; Giroletti et al. 2016). Recently, Schinzel et al. (2015) found 76 new low-frequency associations of γ-ray sources between 5 and 9 GHz by detecting parsec-scale emission through Very Long Baseline Interferometric (VLBI) observations.

We are now investigating 1FHL sources at δ> 0° with the use of high angular resolution VLBI radio observations. There are several goals of this project: supporting the proposed blazar associations by revealing a compact high brightness temperature radio core; searching for new counterparts for UGS; increasing the size of the population of E> 10 GeV sources with high angular resolution observations; and exploring the existence of a correlation between VLBI and E> 10 GeV emission on a sample that is as large and unbiased as possible.

In this paper we present the first results of this project. From the whole 1FHL catalog we extracted a sample of 269 sources in the northern sky with a radio counterpart in the NRAO VLA Sky Survey (NVSS, Condon et al. 1998) within the 95% confidence radius (r95) by using TOPCAT software (Taylor 2005). We call this the 1FHL-n sample. For 185 out of these 269 sources Very Long Baseline Array (VLBA) archival observations are already available. Eighty-four sources (~31%) have never been observed with the VLBA, and 21 of them are UGS. We focus our attention on this these 84 1FHL sources (see Table A.1). For the sources with an association we aim to confirm their low-frequency association and to study their parsec scale properties. For the UGS we want to investigate their nature, possibly detecting a compact radio source associated with them.

The paper is organized as follows: in Sect. 2 we present the observations and data reduction procedures; we show the results in Sect. 3; we discuss and summarize the results in Sect. 4 and in Sect. 5, respectively. In a following paper, we will present a detailed statistical analysis of the entire 1FHL sample, based on new and archival data.

Throughout the paper, we use a ΛCDM cosmology with h = 0.71, Ωm = 0.27, and ΩΛ = 0.73 (Komatsu et al. 2011). The radio spectral index is defined such that S(ν) ∝ να and the γ-ray photon index Γ such that dNphoton/ dEE− Γ.

Table 1

Observations details.

thumbnail Fig. 1

5 GHz VLBA image of the radio counterparts of the γ-ray sources 1FHL J0043.7+3425 (left panel), classified as a FSRQ, and 1FHL J0648.9+1516 (right panel), classified as a BL Lac. The beam size is 1.6 mas × 4.1 mas and 1.3 mas × 3.2 mas, respectively. Levels are drawn at (− 1,1,2,4...) × the lowest contour (i.e., at 1.2 mJy/beam for the source 1FHL J0043.7+3425 and at 1.4 mJy/beam for the source 1FHL J0648.9+1516) in steps of 2. The noise level is 0.31 mJy and 0.37 mJy, respectively.

Open with DEXTER

2. Observations and data reduction

We performed the VLBA observations between 2013 September 30 and December 7, and two additional observing epochs between 2015 July 10 and December 8. We divided the sample into eight observing blocks (Table 1). The observations were carried out at a central frequency of 5 GHz in full polarization; the total bandwidth of 256 MHz was divided into eight 32 MHz sub-bands in each polarization, with a recording rate of 2 Gbps. Observations were carried out in phase reference mode; the duty cycle of 5 min (4 min on target and 1 min on calibrator) was repeated six times, resulting in a net observing time of ~24 min per source. The total observing time for this project is 48 h. As shown in Table 1, during some observing epoch some antennas did not work properly because of technical problems.

Even if the angular resolution at E> 10 GeV (<0.15°) is lower than at lower γ-ray energies (<3.5° at E> 100 MeV), it still is much higher than the typical field of view of VLBI observations (milliarcsecond scale). Moreover, there are 18 sources for which there is not a single well-identified radio counterpart, but rather a set of a few possible candidates (between 2 and 5). For these sources we exploit the capability of the DiFX correlator (Deller et al. 2011) to correlate the data at multiple phase centers, one for each candidate radio counterpart. Thanks to this feature it is possible to observe all of the radio sources within the r95 region with one single pointing. However, some of them are located at a distance from the pointing center that is comparable to the 25 m single-dish primary beam at 5 GHz (~8 arcmin).

We carried out a full calibration by using the software package Astronomical Image Processing System (AIPS, Greisen 2003) with the following steps: phase calibration by applying ionospheric and Earth orientation parameters correction; amplitude calibration based on gain curves and measured system temperatures; parallactic angle correction; removal of residual instrumental phase and delay offset by using the pulse-cal tones table; global fringe fitting on all of the calibrator sources; and transfer of the solutions to the targets.

After this, we split the phase-referenced calibrated visibility data by averaging data in frequency within each sub-band, but not across IFs, and without any time average (the correlator integration time was 4 s). This provided a nominal field of view of 6.2 arcsec. With these single-source datasets, we started the editing and imaging process in DIFMAP (Shepherd 1997). Since the phase tracking centers were based on NVSS observations, which for faint sources can be as inaccurate as a few arcseconds, we often had to image wide sky regions before finding a significant source in the image plane. We determined the accurate positions and shift from the phase tracking center for all of the detected sources (see Table A.1). We then went back to the multi file in AIPS and for each source we corrected the position by using the task CLCOR.

After this calibration procedure, for the six sources (1FHL J0338.4+1304, 1FHL J0516.4+7351, 1FHL J1244.9+5708, 1FHL J1315.0+2346, 1FHL J1942.8+1034, and 1FHL J2223.4+0104) whose calibrators have some extended structures, we imaged the calibrator in DIFMAP and we produced a model that we use in AIPS for two self-calibration cycles with the task CALIB. The source 1FHL J0930.4+8611 is quite strong (S> 200 mJy/beam) and we performed a global fringe-fit on the source itself.

After splitting the single-source visibility datasets, we started a final cycle of imaging and self calibration in AIPS, obtaining for each source the peak flux density and the deconvolved angular size by using the task JMFIT. Finally, for the sources with a significant offset from the pointing center, we corrected for primary beam attenuation by using the widefield VLBI calculator1.

For 11 sources (marked with an asterisk in Table A.1) we observed the NVSS source, which is closer to the 1FHL source centroid but is not coincident with the proposed low-frequency association in the 1FHL. For this reason we exclude these 11 sources from our analysis. We also exclude the supernova remnant source 1FHL J1911.0+0905 and the final sample discussed in this paper consists of 72 sources.

3. Results

3.1. VLBI properties and detection rate

We detect radio emission on VLBI scales for 67/72 sources, which corresponds to an overall detection rate of 93%. For blazars the detection rate is 100% (51/51), while for the UGS it is 76% (16/21). The significance level for the detection of these sources is always 5σ. The VLBI coordinates, the 10500 GeV energy flux (Sγ), the 1.4 GHz NVSS flux density (S1.4,NVSS), the 5 GHz VLBA flux density (S5,VLBI), and the brightness temperature of the detected sources are listed in Table A.2.

In Fig. 1 we show an example of 5 GHz images of the FSRQ source 1FHL J0043.7+3425 (left panel), and the BL Lac source 1FHL J0648.9+1516 (right panel), which are representative of the global dataset. For these 67 sources we reveal for the first time a VLBI compact structure. The detection is also confirmed by a visual inspection of the visibility data.

In five sources (1FHL J0338.4+1304, 1FHL J1315.0+2346, 1FHL J1809.3+2040, 1FHL J1942.8+1034 and 1FHL J2223.4+0104) we reveal some extended structures close to the compact region. Examples of extended sources are shown in Fig. 2. We used the AIPS task JMFIT to fit the brightness distribution of each extended source in the image plane with two elliptical Gaussian components. In Table 2 we list the full widths at half maximum (FWHM) of the major and minor axes of the deconvolved elliptical Gaussian component measured in mas.

thumbnail Fig. 2

5 GHz VLBA image of the radio counterparts of the γ-ray sources 1FHL J1315.0+2346 (left panel), classified as a BL Lac, and 1FHL J1942.8+1034 (right panel), classified as an active galaxy of uncertain type. The beam size is 1.2 mas × 3.0 mas and 1.3 mas × 3.1 mas, respectively. Levels are drawn at (− 1,1,2,4...) × the lowest contour (i.e., at 0.4 mJy/beam for the source 1FHL J1315.0+2346 and at 2.0 mJy/beam for the source 1FHL J1942.8+1034) in steps of 2. The noise level is 0.12 mJy and 0.50 mJy, respectively.

Open with DEXTER

thumbnail Fig. 3

Flux density distribution for the detected sources. Top panel: 1.4 GHz NVSS flux density distribution; middle panel: 5 GHz VLBA flux density distribution; bottom panel: distribution of the ratio between the 5 GHz VLBA and the 1.4 GHz NVSS flux densities. Solid red line: All sources; green dashed line: Blazars; blue dashed line: UGS.

Open with DEXTER

Table 2

Modelfit details for the extended sources.

3.2. Flux density distribution

The distribution of S1.4,NVSS of the newly VLBA detected sources is shown in the top panel of Fig. 3 (red solid line), which has a median value of 40.6 mJy. In the middle panel of Fig. 3 we show the S5,VLBI distribution (red solid line), which has a median value of 16.3 mJy. In both panels we also show separately the distribution of flux densities for the blazars (green dashed line) and for the UGS (blue dashed line). We note that for the UGS the median values for both S1.4,NVSS (31.2 mJy) and S5,VLBI (8.9 mJy) are lower than for the blazars (median values of 47.3 mJy and 19.5 mJy for S1.4,NVSS and S5,VLBI, respectively).

In the bottom panel of Fig. 3 we show the distribution of the logarithm of the ratio between the flux density on parsec scale (S5,VLBI) and on kiloparsec scale (S1.4,NVSS). The distribution peaks at ~0.5, indicating that for most sources there is a considerable amount of radio emission on intermediate scales between the parsec and kiloparsec scales.

By an inspection of both NVSS and VLBI flux densities, it emerges that there are some interesting sources. In ~10% of the sources the ratio S5,VLBI/S1.4,NVSS is significantly higher than 1, as in the case of the UGS 1FHL J0307.4+4915, with S5,VLBI = 300 ± 30 mJy, while S1.4,NVSS is 56.0 ± 1.7 mJy. At the other end of the distribution there are sources for which the ratio S5,VLBI/S1.4,NVSS is significantly lower than 1. For example, for the sources 1FHL J0333.6+2918, 1FHL J0334.0+6539, 1FHL J0601.0+3838, and 1FHL J2127.8+3614, the ratio S5,VLBI/S1.4,NVSS is between 0.1 and 0.2.

The immediate interpretation is that the sources with S1.4,NVSSS5,VLBI have extended emission (on arcsecond scale), while in the case of S5,VLBIS1.4,NVSS some degree of variability could be responsible of the higher observed VLBI flux density. However, it is important to stress that we are comparing non-concurrent observations at different spatial scales and at different frequencies.

3.3. High energy properties

The distribution of 1FHL Sγ for the detected sources is presented in the left panel of Fig. 4 (red solid line), showing a median value of 4.8 × 10-12 erg cm-2 s-1. In particular, blazars (green dashed lines) and the UGS (blue dashed lines) have median values of 5.1 × 10-12 and 4.5 × 10-12 erg cm-2 s-1, respectively.

The photon index distribution is shown in the right panel of Fig. 4, and has a median value of 2.3 both for the blazars (green dashed lines) and the UGS (blue dashed lines).

3.4. Proposed counterparts for unassociated sources

For 16 out of the 21 UGS studied in this paper, we detect radio emission on VLBI scales and we propose a possible radio NVSS counterpart. By using the approach proposed by Schinzel et al. (2015), we determine the likelihood ratio for the association of each of the detected UGS. Schinzel et al. (2015) propose a likelihood ratio threshold value of 8, chosen from the distribution of the obtained values of their full sample, to claim the association of the γ-ray source with the proposed radio counterpart. For all of the detected UGS in the present work but two (1FHL J0605.4+2726 and 1FHL J1212.2+2318), the likelihood ratio of their association is higher than 8. The detected UGS, along with further details about other γ-ray associations, are listed in Table 3.

We note that 8 out of the 16 detected UGS in our sample are included in the all-sky radio survey, between 5 and 9 GHz, of sky areas surrounding all of the second Fermi-LAT catalog (2FGL, Nolan et al. 2012) UGS by Schinzel et al. (2015). Our proposed low-frequency associations always match those proposed by Schinzel et al. (2015) and by using their 7 GHz VLBA observations we obtain the spectral index values between 5 and 7 GHz (see Table 3).

For these eight 1FHL UGS included in the 2FGL catalog, our proposed low-frequency counterparts are in agreement with those proposed in other works which make use of complementary association methods based on the analysis of their multiwavelength (MWL) properties, such as the investigation of infrared colors and the low-frequency spectral index (Massaro et al. 2013a,b, 2014, 2015). Our proposed counterparts for the sources 1FHL J0644.2+6036, 1FHL J1223.3+7953, and 1FHL J0601.0+3838 are compatible with the X-ray associations proposed by Landi et al. (2015) and Paggi et al. (2013), as indicated in the notes in Table 3.

thumbnail Fig. 4

Distribution of γ-ray energy flux (left panel) and photon index (right panel) above 10 GeV for the detected sources. Solid red line: all sources; green dashed line: Blazars; blue dashed line: UGS.

Open with DEXTER

In Table 3, we also list the sources 1FHL J0332.1+6307, 1FHL J0338.4+1304, 1FHL J0425.3+6320, 1FHL J0601.0+3838, 1FHL J0650.4+2056, 1FHL J1841.1+2914, 1FHL J2002.6+6303, and 1FHL J2004.7+7003, for which new low-frequency counterparts are proposed in the 3FGL catalog. For the source 1FHL J0307.4+4915 a new low-frequency counterpart is now proposed in the second catalog of hard Fermi-LAT sources (2FHL, Ackermann et al. 2016). All of these new associations are coincident with those proposed in this work.

For the sources 1FHL J0053.9+4030 and 1FHL J1619.8+7540 we propose, for the first time, a new low-frequency association. We note that the source 1FHL J0053.9+4030 is one of the faintest sources (SVLBI ~ 5 mJy) in our sample and it is not included in any other Fermi catalog. For the sources 1FHL J0605.4+2726 and 1FHL J1212.2+2318, which are not included in any other Fermi catalog, we propose a new low-frequency counterpart, but for their association we find a likelihood ratio lower than 8. In both cases the proposed radio counterpart is located very close to the edge of the 1FHL error ellipse.

Among the five undetected UGS, the sources 1FHL J1406.4+1646 and 1FHL J0828.9+0902 are now associated in the 3FGL with sources not included among our observed phase centers; the source 1FHL J0625.9+0002 is included in the 3FGL and classified as UGS; the sources 1FHL J0625.9+0002 and 1FHL J0432.2+5555 are not included in any other Fermi catalog.

Table 3

Details of the 16 detected UGS.

3.5. Measurements of the brightness temperature

The high resolution of VLBI observations allows a direct measurement of the brightness temperature of a source. However, because blazar jets are highly relativistic, and therefore Doppler boosted, it is difficult to measure the intrinsic brightness temperature (). What we measure is the observed brightness temperature (), which is linked to by the relation , where δ is the Doppler factor and α is the spectral index.

By fitting the brightness distribution of each source in the image plane with elliptical Gaussian components, and assuming α = 0, we determine for each source with the equation (Piner et al. 1999; Tingay et al. 1998)

where Score corresponds to the fitted core flux density at 5 GHz measured in Jy, a and b are the FWHM of the major and minor axes of the elliptical Gaussian core component measured in mas, z is the redshift, and ν is the observing frequency in GHz.

For the sources without any redshift estimation we assume z = 0.2, which corresponds to the median redshift value of 1FHL BL Lac objects. We note that for some sources the radio core is only slightly resolved or is unresolved (i.e., less than half of the beam size).

For those sources we determine measurements as upper limits to a and b based on the Gaussian deconvolved size reported by the AIPS task JMFIT (Greisen 2003). When a nominal value is provided we use it as measurement, when only a maximum value is reported we treat it as an upper limit, resulting in a lower limit. In a few cases, also the maximum value reported by JMFIT is 0; for these sources we conservatively set the upper limit to the minimum size value found among the other sources.

The resulting brightness temperature values are reported in Table A.2. For the 36 sources whose radio core is resolved, the average is on the order of 2 × 1010 K, which is close to the expected value for equipartition. Only five sources have varying in the range 7 × 10101 × 1011 K.

4. Discussion

4.1. General characterization of the observed sample

The 1FHL catalog provides us, for the first time, with a deep, large and unbiased sample of sources detected in the energy range 10500 GeV. It fills the gap existing between the HE and VHE regimes, represented by 3FGL and TeVCat2, respectively.

An important step on the basis of the analysis and interpretation of the physical properties of these sources is the investigation and comparison of the multifrequency properties for the entire catalog. However, such a study cannot be carried out while neglecting a significant fraction of the sample, i.e., the 84 sources without VLBI data presented here. These 84 sources do not represent a random 1FHL subsample. Their 1.4 GHz NVSS radio flux density distribution is significantly different with respect to the other sources belonging to the 1FHL-n sample (Fig. 5). A KS-test provides a probability of the two distributions being drawn from the same population of ~2 × 10-12. Another peculiarity of our sample is that 61% of the selected AGNs are of HSP type, while in the complementary sample of 1FHL-n sources already observed with VLBI the fraction of HSPs is 40%. This occurs mainly because we are targeting a 1FHL subsample of sources with the faintest radio flux densities, in agreement with the blazar sequence trend (e.g., Fossati et al. 1998; Ghisellini et al. 1998). Moreover, the fraction of HSP objects without a known redshift in our sample is 65%, while in the 1FHL-n they represent 42%. Finally, we note that while 10% of 1FHL sources at Dec > 0° are classified as UGS, in our sample the fraction of UGS is 29%. We are targeting 88% (21/24) of the 1FHL UGS in the northern sky with at least one NVSS counterpart within the r95. Therefore, we are not only completing the VLBI data collection for the whole sample of hard γ-ray sources, but we are also characterizing an unexplored population.

thumbnail Fig. 5

Histogram of the NVSS flux density distribution for VLBA observed sources (solid line) and unobserved sources (dashed line) in our sample.

Open with DEXTER

4.2. VLBI properties and unassociated sources

The vast majority of the detected sources in our sample are radio weak (with flux densities of a few mJy). For all of the detected sources we reveal a VLBI compact component and we confirm their blazar nature. This finding provides an independent validation of other association methods based on the analysis of their multiwavelength properties, such as the investigation of infrared colors, the low-frequency spectral index, and the X-ray association (e.g., D’Abrusco et al. 2013; Massaro et al. 2013a; Nori et al. 2014; Landi et al. 2015).

It is interesting to note that the detection rate of compact components for the UGS is remarkably high (76%). This is a first indication that the nature of the UGS is similar to the associated sources, i.e., they are mostly blazars. This hypothesis is supported by the fact that both S1.4,NVSS and S5,VLBI for the blazars and the UGS are distributed in a similar way (see Fig. 3). This also shows how much the NVSS survey is useful for a pre-selection of compact radio counterparts, and deeper Jansky very large array and Australia telescope compact array observations are only needed in special cases (Schinzel et al. 2015). However, the VLBI observations are necessary for the actual selection of the right association.

We note that for the UGS both the NVSS and VLBI flux densities are lower than for the associated sources (Fig. 3); therefore, they may be weaker blazars not yet identified. Moreover, for eight detected UGS we calculated the spectral index values between 5 and 7 GHz VLBA data, which are either flat or inverted, indicating the presence of a self-absorbed region. In particular for the source 1FHL J0644.2+6036 the spectral index is strongly inverted (α = −2.7), exceeding the canonical value of 2.5 for a homogeneous synchrotron emission region. This is an indication that variability is occurring in this source. A clear example of flux density variability is provided by the source 1FHL J0307.4+4915, for which we find S5,VLBI = 300 ± 30 mJy, while S1.4,NVSS is 56.0 ± 1.7 mJy.

4.3. Brightness temperature

For the detected sources we estimated the VLBI core brightness temperature, which can be used to distinguish between the physical processes occurring in the source. In a synchrotron source there are two main physical mechanisms which can limit the intrinsic brightness temperature. One is the inverse Compton (IC) process, which limits in the range ~5 × 1011−1012 K (Kellermann & Pauliny-Toth 1969). When this limit is exceeded the IC process leads to rapid and catastrophic electron energy losses. The second relevant physical mechanism is the energy equipartition between particles and magnetic field which limits the in the range ~5 × 1010−1011 K (Readhead 1994).

For the vast majority of the sources in our sample whose cores are resolved we obtain brightness temperature values on the order of 2 × 1010 K, which is close to the expected value for equipartition. Therefore, we expect the energy to be almost equally stored in the magnetic field and in the radiating particles. These values are consistent with the typical HSP blazar , and are in agreement with the values obtained in other works (e.g., Piner et al. 2010; Lico et al. 2012, 2014; Piner & Edwards 2014). From these values no high Doppler factors are required and there is no evidence of a strong beaming. However, we note that at 5 GHz we are observing these sources in their self-absorbed part of the synchrotron spectrum, therefore the flux densities and the brightness temperatures may be underestimated.

5. Summary and conclusions

In the present work we targeted and characterized the 1FHL sources in the northern sky with the faintest radio flux densities; these sources represent an important extension to an unexplored region of the parameter space, and are essential to obtain a sample that is as large and unbiased as possible to explore the possible correlation between radio VLBI and E> 10 GeV emission.

  • For all of the detected sources we reveal a compact high brightness temperature VLBI radio core and we confirm their blazar nature. The vast majority are radio weak sources, with a VLBI flux density median value of 16.3 mJy, consistent with the predictions of the blazar sequence (e.g., Fossati et al. 1998; Ghisellini et al. 1998).

  • Thanks to the new VLBI observations we propose new low-frequency counterparts for 16 1FHL UGS. We note that for 12 out of the 16 detected UGS our proposed low-frequency counterparts are in agreement with those proposed by using indirect and complementary association methods based on the analysis of their MWL properties, and with those included in the 2FHL and 3FGL catalogs. For the remaining four 1FHL UGS we propose a low-frequency counterpart for the first time.

  • For the sources whose radio core is resolved we obtain brightness temperature values on the order of 2 × 1010 K, which are close to the value expected for the equipartition of energy between particles and magnetic field. Therefore, there is no evidence of a strong beaming and no high Doppler factor values are required. This result increases the statistical significance of the similar brightness temperature values obtained for other HSP sources (e.g., Giroletti et al. 2004; Piner & Edwards 2014).

Ackermann et al. (2011) found a strong and significant correlation between radio and γ rays in the 0.1100 GeV energy range for blazars belonging to the first catalog of AGNs detected by Fermi (1LAC, Abdo et al. 2010b); however, no systematic investigation for the existence of a possible correlation between radio and VHE γ rays has been made. The main reason is that at VHE there is no homogeneous, deep and large survey that allows us to perform a similar correlation analysis; the IACTs have small field of view and they mainly observe in pointing mode. All of these limitations will be overcome in the coming years thanks to the new generation Cherenkov Telescope Array (CTA, Dubus et al. 2013). The investigation in the VHE domain is important because it provides us with details about the blazar sequence (i.e., the anti-correlation between the synchrotron luminosity and the SED peak frequency (Fossati et al. 1998)), and the interaction of VHE photons with the EBL (e.g., Ackermann et al. 2012).

In a following paper, by using the VLBI flux densities presented in this work complemented by VLBI archival observations, we will explore and quantify the possible radio-VHE correlation for the 1FHL AGN sample.


Acknowledgments

We thank Leonid Petrov for his support and the fruitful discussions. We thank the anonymous referee for the helpful comments and discussions which improved the manuscript. This work is based on observations obtained through the S6340 VLBA project, in the framework of the Fermi-NRAO cooperative agreement. The VLBA makes use of the Swinburne University of Technology software correlator, developed as part of the Australian Major National Research Facilities Programme and operated under licence (Deller et al. 2011). We acknowledge financial contribution from grant PRIN-INAF-2011 and PRIN-INAF-2014. This research has made use of NASA’s Astrophysics Data System, of the VizieR catalog access tool, CDS, Strasbourg, France and the TOPCAT software (Taylor 2005). The National Radio Astronomy Observatory is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc. For this paper we made use of the NASA/IPAC Extragalactic Database NED, which is operated by the JPL, Californian Institute of Technology, under contract with the National Aeronautics and Space Administration. The Fermi-LAT Collaboration acknowledges support for LAT development, operation, and data analysis from NASA and DOE (United States), CEA/Irfu and IN2P3/CNRS (France), ASI and INFN (Italy), MEXT, KEK, and JAXA (Japan), and the K. A. Wallenberg Foundation, the Swedish Research Council, and the National Space Board (Sweden). Science analysis support in the operations phase from INAF (Italy) and CNES (France) is also gratefully acknowledged.

References

Appendix A: Additional tables

Table A.1

Sample source details.

Table A.2

Details of the 67 detected sources.

All Tables

Table 1

Observations details.

Table 2

Modelfit details for the extended sources.

Table 3

Details of the 16 detected UGS.

Table A.1

Sample source details.

Table A.2

Details of the 67 detected sources.

All Figures

thumbnail Fig. 1

5 GHz VLBA image of the radio counterparts of the γ-ray sources 1FHL J0043.7+3425 (left panel), classified as a FSRQ, and 1FHL J0648.9+1516 (right panel), classified as a BL Lac. The beam size is 1.6 mas × 4.1 mas and 1.3 mas × 3.2 mas, respectively. Levels are drawn at (− 1,1,2,4...) × the lowest contour (i.e., at 1.2 mJy/beam for the source 1FHL J0043.7+3425 and at 1.4 mJy/beam for the source 1FHL J0648.9+1516) in steps of 2. The noise level is 0.31 mJy and 0.37 mJy, respectively.

Open with DEXTER
In the text
thumbnail Fig. 2

5 GHz VLBA image of the radio counterparts of the γ-ray sources 1FHL J1315.0+2346 (left panel), classified as a BL Lac, and 1FHL J1942.8+1034 (right panel), classified as an active galaxy of uncertain type. The beam size is 1.2 mas × 3.0 mas and 1.3 mas × 3.1 mas, respectively. Levels are drawn at (− 1,1,2,4...) × the lowest contour (i.e., at 0.4 mJy/beam for the source 1FHL J1315.0+2346 and at 2.0 mJy/beam for the source 1FHL J1942.8+1034) in steps of 2. The noise level is 0.12 mJy and 0.50 mJy, respectively.

Open with DEXTER
In the text
thumbnail Fig. 3

Flux density distribution for the detected sources. Top panel: 1.4 GHz NVSS flux density distribution; middle panel: 5 GHz VLBA flux density distribution; bottom panel: distribution of the ratio between the 5 GHz VLBA and the 1.4 GHz NVSS flux densities. Solid red line: All sources; green dashed line: Blazars; blue dashed line: UGS.

Open with DEXTER
In the text
thumbnail Fig. 4

Distribution of γ-ray energy flux (left panel) and photon index (right panel) above 10 GeV for the detected sources. Solid red line: all sources; green dashed line: Blazars; blue dashed line: UGS.

Open with DEXTER
In the text
thumbnail Fig. 5

Histogram of the NVSS flux density distribution for VLBA observed sources (solid line) and unobserved sources (dashed line) in our sample.

Open with DEXTER
In the text

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

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

Initial download of the metrics may take a while.