Identifying the 3FHL Catalog. II. Results of the KOSMOS Optical Spectroscopy Campaign

The Large Area Telescope (LAT; Atwood et al. 2009), mounted on the Fermi Gamma-ray Space Telescope, represented a breakthrough for the study and understanding of the high-energy universe, and in particular of blazars, the most extreme class of active galactic nuclei and the most numerous population detected in γ-rays. The Third Fermi-LAT Catalog of High-Energy Sources (3FHL Ajello et al. 2017), based on seven years of Fermi observations, is the latest catalog of sources detected at >10 GeV and fully exploits the improvement in performances delivered by "Pass 8," the newest event-level analysis method (Atwood et al. 2013). The 3FHL reports the detection of more than 1500 sources all-sky, a dramatic improvement of a factor of 3 over its predecessor, the 1FHL¹ catalog (Ackermann et al. 2016). The 3FHL represents the current deepest look at the very high-energy sky and in all likelihood it will be used to plan most of the observations of the upcoming Cherenkov Telescope Array (CTA; Cherenkov Telescope Array Consortium et al. 2017).

To maximize the scientific impact of the 3FHL catalog, the sample should be completely identified (i.e., the nature of all sources should be known) and the vast majority of the objects should have a redshift (z). At the present day, however, while the 3FHL catalog contains more than 1500 sources detected at >10 GeV, 1227 of which are of confirmed extragalactic origin, only a minority of them (42.5%) have a measured z and a significant fraction (∼25%) lacks both a redshift and a secure classification. Particularly, the 3FHL contains 290 blazar candidates of uncertain type (BCUs) and 177 unassociated sources. Spectroscopic observations of most of these sources are needed to fully realize the scientific potential of the 3FHL catalog. The redshift measurement is particularly important, since it allows us to use the 3FHL catalog to investigate a variety of topics, such as the blazars energetics and emission mechanism (see, e.g., Ghisellini et al. 2017); the extragalactic background light (EBL), i.e., the integrated emission of all stars and galaxies in the universe, and its evolution with redshift (Abramowski et al. 2012; Ackermann et al. 2012; Domínguez et al. 2013); the role of blazars in accelerating cosmic rays (see, e.g., Furniss et al. 2013); to study the evolution of blazars across cosmic time (Ajello et al. 2014).

At the present day, spectroscopic follow up of unclassified blazars have been mainly performed with 4 m or 8 m class telescopes (e.g., Sbarufatti et al. 2005, 2006; Shaw et al. 2013; Massaro et al. 2014; Paggi et al. 2014; Landoni et al. 2015; Ricci et al. 2015; Álvarez Crespo et al. 2016a, 2016b; Marchesini et al. 2016; Peña-Herazo et al. 2017). These works showed that 4 m telescopes are effective in distinguish between flat spectrum radio quasars (FSRQs) and BL Lacs. FSRQs present strong emission features and their redshift are therefore easily measured; BL Lacs, instead, are typically characterized by narrow (equivalent width, EW ≤ 5–20 Å; see, e.g., Marcha et al. 1996) emission lines, although this EW-based classification is phenomenological, rather than physical, and several cases of "transition blazars," i.e., objects with significant EW variability, mostly BL Lacs showing broad permitted emission lines, are reported in the literature (see, e.g., Ulrich 1981; Corbett et al. 2000; Ghisellini et al. 2011; Shaw et al. 2012; Ruan et al. 2014). Furthermore, BL Lacs can be divided in three different classes, based on the frequency of the synchrotron peak: sources with ${\mathrm{log}}_{10}({\nu }_{\mathrm{peak}}^{{\rm{S}}}/\mathrm{Hz})\lt 14$ are classified as low-synchrotron peak (LSP), sources with $14\,\lt {\mathrm{log}}_{10}({\nu }_{\mathrm{peak}}^{{\rm{S}}}/\mathrm{Hz})\,\lt 15$ are classified as intermediate-synchrotron peak (ISP) and sources with ${\mathrm{log}}_{10}({\nu }_{\mathrm{peak}}^{{\rm{S}}}/\mathrm{Hz})\gt 15$ are classified as high-synchrotron peak (HSP). Historically, it has been possible to obtain a spectroscopic redshift for 67%, 37% and 48% of LSP (including FSRQs), ISP and HSP objects, respectively (Shaw et al. 2012, 2013).

In this paper, we report the results of a spectroscopic follow-up campaign with the 4 m Mayall telescope at the Kitt Peak National Observatory (KPNO) of 28 3FHL sources selected among those lacking of redshift information. The work is organized as follows: in Section 2, we present the sample selection criteria. In Section 3, we describe the observing process and the data analysis. In Section 4, we report the results of the spectroscopic campaign, both in general terms and in detail for each of the 28 sources. Finally, we report our conclusions in Section 5.

Our long-term goal is to provide redshifts and type information for all the 706 sources with no redshift or classification in the 3FHL catalog. In this work, we report the results of the first step of this ambitious campaign, i.e., the observation of a subsample of 28 sources with the Mayall 4 m telescope at KPNO.² Out of this 28 sources, 23 are blazar candidate of uncertain type (BCUs; Ackermann et al. 2015). BCUs have multiwavelength properties which allow to reliably classify them as blazars; however, their low-quality optical spectra prevent us to distinguish their class, i.e., either BL Lacs or FSRQs.

The other five objects are instead classified as "unassociated", i.e., in 3FHL catalog they are not associated with a counterpart. These objects are part of a larger sample that we will present in a companion paper (A. Kaur et al. 2019, in preparation). In this work, we look for potential 0.5–10 keV counterparts of the 3FHL sources using archival Swift-XRT observations: the XRT position accuracy of a few arcseconds allows one to easily identify the optical and/or radio counterparts for these objects. In Figure 1, we report the Swift-XRT 0.3–10 keV images of the five unassociated objects in our sample: as can be seen, four out of five 3FHL sources have a single, bright (count rate in the 0.5–10 keV band r ≳ 0.01 cts s⁻¹)) counterpart within the LAT 95% confidence positional uncertainty. The only exception is 3FHL J2104.5+2117, where no significant 0.3–10 keV emission is observed within the LAT source region. However, an X-ray source associated with a radio counterpart is located at ∼45 from the LAT centroid, and we thus choose this object (radio source NVSS J210415+211805) as the most likely counterpart.

Zoom In Zoom Out Reset image size

Besides their lacking of redshift and type information, we selected the sources in our sample with the following criteria:

• 1.  
  Having available optical magnitude, and the magnitude being V ≤ 21.5.
• 2.  
  Being bright in the hard γ-rays (f_{50–150 GeV} > 10⁻¹³ erg s⁻¹ cm⁻²). Selecting 3FHL objects bright in the 50–150 GeV band ensures that the completeness of the 3FHL catalog evolves to lower fluxes as more optical observations are performed.
• 3.  
  Being observable from Kitt Peak (decl. > −20°) and during our observations (in August, this implies R.A. ≥ 17^h00^m00^s and R.A. ≤ 3^h00^m00^s; in October, it is R.A. ≥ 18^h00^m00^s and R.A. ≤ 7^h00^m00^s).

66 3FHL sources satisfy all these criteria: 61 BCUs and five "unassociated" sources studied in A. Kaur et al. (2019, in preparation). Our 28 sources were selected among these 66 objects with the goal of covering a wide range of optical magnitudes (V = [14–20]) and, consequently, of potential redshifts and luminosities. We report in Table 1 a summary of the sources analyzed in this work.

All the spectra analyzed in this work were acquired with the KOSMOS spectrograph mounted on the KPNO Mayall 4 m (Martini et al. 2014), using the Blue VPH grism in conjunction with the 3500–6200 Å, 09 slit. This experimental setup corresponds to a dispersion of ∼4 Å pixel⁻¹ and a spectral resolution R ∼ 2100. The data were taken with the slit aligned along the parallactic angle. The observations were performed in seven "grey" nights of observations, three nights in 2017 August and four in 2017 October; however, due to poor weather conditions and technical issues, the effective number of nights of observations is ∼4.5. The observing conditions were spectroscopic. All the observations were performed remotely.

All sources have been observed at least three times³ and the single observations have then been combined to reduce cosmic rays contribution and instrumental effects. The data reduction and spectral extraction procedure has been carried out using a standard IRAF pipeline (Tody 1986), with bias subtraction, flat field normalization and bad pixel correction. All spectra have also been visually inspected to remove any artificial feature.

The spectra have been wavelength calibrated using Iron-Argon lamps, whose spectra were acquired either before or after each source observation, to take into account possible shifts in the pixel-λ calibration, due to changes in the telescope position during the night. The spectra have then been flux calibrated using a spectrophotometric standard: in each observing night we observed two spectrophotometric standards, one at the beginning and the other at the end of the night: the standards were observed using the same 09 slit used in the rest of the analysis. Finally, each spectrum has been corrected for Galactic reddening, using the Cardelli et al. (1989) extinction law and the E(B − V) value obtained from the NASA/IPAC Infrared Science Archive online tool⁴ : we adopted the measurements of Schlafly & Finkbeiner (2011).

We report in Table 2 a summary of the spectral analysis results. The spectra signal-to-noise ratios (S/Ns) have been computed on the normalized spectra. The S/N has then been measured in five regions, each of which having size Δλ ∼ 50 Å, chosen in featureless parts of the spectrum in the energy range 4500–6000 Å.

The flux-calibrated spectra, together with the normalized spectra, are shown in Figure 2.⁵ The normalized spectra are obtained dividing the flux-calibrated spectrum by a power law fit of the continuum.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

All spectra have been visually inspected to identify candidate emission and absorption features. In order to deem a feature reliable, we verified its existence in each of the individual spectra which are combined in the final one.

As a first result of this follow-up program, we are able to classify all the sources in our sample: 27 out of 28 objects are BL Lacs, while the remaining one is a FSRQ. We determine a redshift for three of these objects and a lower limit on z for other four sources.

Finally, 21 objects are classified as featureless BL Lacs. These are sources where a 4 m instrument does not allow to reach an S/N high enough to observe faint features, but represent a potential candidate for further observations with 8 m or 10 m telescopes. Notably, while we find no clear feature in 75% of the objects in our sample, in a recent work by (Paiano et al. 2017a) only 4 out of 20 blazars (20% of their sample) observed with the 10 m Gran Telescopio Canarias (GTC) are found to be featureless.

4.1. Comments on Individual Sources

In this work, we report the results of the first part of an optical spectroscopy campaign aimed at increasing the spectral completeness of the 3FHL catalog. We observed 28 3FHL sources with the KOSMOS spectrograph mounted on the 4 m Mayall telescope at Kitt Peak: 23 of these objects are BCUs, i.e., blazars of uncertain classification, while the remaining five are sources unassociated in the 3FHL catalog for which we found a optical-radio counterpart through a follow up with Swift-XRT (A. Kaur et al. 2019, in preparation).

We were able to reliably classify all the 28 sources in our sample: 27 objects are BL Lacs, and only one source is a FSRQ (3FHL J0134.4+2638). The fact that the vast majority of sources in our sample are BL Lacs is not unexpected: among the 873 blazars already classified in the 3FHL catalog, 731 BL Lacs (83.7% of the sample of classified blazars) and 142 FSRQs (16.3%). This is due to the fact that above 10 GeV, i.e., in the energy range covered in the 3FHL catalog, BL Lacs are significantly brighter than FSRQs.

Overall, we placed a constraint on the redshift of 25% of the objects in our sample. This relatively low fraction is consistent with those obtained in the vast majority of works based on 4 m instruments (10%–35%; see, e.g., Landoni et al. 2015; Ricci et al. 2015; Álvarez Crespo et al. 2016a; Peña-Herazo et al. 2017). The fraction of sources for which it is possible to obtain a redshift measurement is instead significantly larger in surveys performed using 8 m+ telescopes (60%–80%; see, e.g., Sbarufatti et al. 2005, 2006; Paiano et al. 2017a). For this reason, we plan to continue our spectroscopic campaign using the 8 m Gemini-N and -S telescopes. We have been granted five nights of observations with these two instruments through Fermi (Cycle 11, proposal ID: 111128, PI S. Marchesi). The observations will take place in 2019 and we plan to observe 40–50 sources.

We acknowledge funding under NASA contract 80NSSC17K0503. We thank Amy Robertson, David Summers and Doug Williams for the help provided during the observing nights, and Marco Landoni for the useful comments on the data reduction and analysis process.

This work made use of data supplied by the UK Swift Science Data Centre at the University of Leicester, as well as of the TOPCAT software (Taylor 2005) for the analysis of data tables.