BASS. XXI. The Data Release 2 Overview

Although active galactic nuclei (AGNs) and the supermassive black holes (SMBHs) that power them have been studied for decades, there are still many key unresolved questions concerning the nature of these systems and how their evolution may be related to the galaxies that host them. While there are some clear SMBH–host correlations, such as the one between SMBH mass and bulge properties (Kormendy & Ho 2013) or the similar redshift evolution of star formation and SMBH growth (e.g., Heckman & Best 2014; Yang et al. 2019), it is not yet clear whether SMBH growth and AGN output (or "feedback," e.g., radiation, winds, jets) affects the host galaxy interstellar medium, star formation, and molecular gas (e.g., Fabian 2012). On smaller scales, it is not yet entirely clear what is the structure of the obscuring torus, what is its connection to its surroundings (e.g., Netzer 2015), and what is the role of obscuration by galaxy-scale gas and nuclear starbursts. Moreover, the stochasticity of SMBH fueling and consequent AGN emission (e.g., Hickox et al. 2014; Schawinski et al. 2015), combined with inherent biases in AGN survey techniques when identifying obscured AGNs (e.g., Hickox & Alexander 2018) make the connections between SBMH growth and feedback even less clear. Therefore, the only way to address the many open questions concerning SMBHs and AGNs, even in the local universe, and to construct statistically significant global trends and/or correlations is to have large samples (e.g., N > 100) surveyed with multiwavelength observations. Extragalactic hard (>2 keV) X-ray surveys provide one of the most complete ways to study growing black holes (BHs) in an unbiased way (see, e.g., Brandt & Alexander 2015; Brandt & Yang 2021). A large fraction, and indeed the majority, of the AGN population is obscured, and therefore the construction of a complete AGN census requires the identification of both obscured and unobscured sources (see review by, e.g., Hickox & Alexander 2018). At even higher energies, the ultrahard X-rays (>10 keV) provide a more complete tracer of the radiation for obscured AGNs (N_H > 10²² cm⁻²) and even some Compton-thick (CT) AGNs (N_H > 10²⁴ cm⁻²; e.g., Ricci et al. 2015; Koss et al. 2016). An all-sky survey in the ultrahard X-ray band (>10 keV) thus provides an important way to answer the fundamental questions of SMBH growth and its links to host galaxy evolution, as well as many open questions concerning AGN physics, for a complete, unbiased sample of AGNs.

Over the past 20 yr, great progress has been made in surveying the ultrahard X-ray sky to increasing depths with the Burst Alert Telescope (BAT; Barthelmy et al. 2005) at 14–195 keV on board the the Neil Gehrels Swift Observatory (Gehrels et al. 2004) and the IBIS instrument (Ubertini et al. 2003) at 17–60 keV on board the INTEGRAL observatory (Winkler et al. 2003). Thanks to its wide field of view (FOV; 1.4 sr half coded), BAT monitors roughly 80% of the sky every day, providing regularly sampled average emission properties of objects. INTEGRAL/IBIS, with a roughly 13 times smaller FOV (0.11 sr half coded) but better angular resolution (12${}^{{\prime} }$ versus 195 for BAT), has focused on targeting the Galactic plane and particularly the Galactic center region. Thus, the Swift BAT survey provides a uniform all-sky census of the average ultrahard X-ray emission of AGNs. Early analysis (e.g., Markwardt et al. 2005; Tueller et al. 2008) of the (ongoing) all-sky survey with BAT found that the brightest ultra-hard-X-ray-selected AGNs in the sky can probe nearby (z < 0.1) AGNs, including highly obscured systems; low-redshift, high-luminosity AGNs and quasars (0.05 ≲ z ≲ 0.3); and much more distant, beamed AGNs (reaching out to z ≳ 3). Specifically, the unbeamed BAT-detected AGNs span the moderate-to-high-luminosity end of the X-ray luminosity function (XLF; e.g., Maccacaro et al. 1991; Comastri et al. 1995; Gilli et al. 2007; Ueda et al. 2014; Aird et al. 2015; Buchner et al. 2015; Ananna et al. 2019) at all accessible redshifts, thus providing critical low-redshift templates for high-z sources detected in much deeper, pencil-beamed surveys (e.g., CDF-S, Luo et al. 2017; COSMOS, Civano et al. 2016). The low-redshift regime of the BAT AGN survey thus strongly complements high-redshift AGN surveys, providing a legacy data set with both high spatial resolution ³⁶ and very high signal-to-noise ratio data obtained through relatively short observations. Finally, some AGNs are uniquely identified as AGNs in the hard X-rays (e.g., Smith et al. 2014). Other all-sky selection methods, such as WISE mid-IR (MIR) colors, are only able to uniformly classify the most luminous AGNs (${L}_{2\mbox{--}10\ \mathrm{keV}}^{\mathrm{obs}}$ > 10⁴⁴ erg s⁻¹; e.g., Ichikawa et al. 2017) owing to contamination from star formation in the host galaxies. Moreover, methods that use strong emission-line ratio diagnostics to identify AGNs (e.g., "BPT" Baldwin et al. 1981; Veilleux & Osterbrock 1987; Kewley et al. 2006) only select about half of the ultra-hard-X-ray-selected BAT AGNs (Koss et al. 2017). This is likely due to the contribution of star-forming regions in the AGN host galaxies (e.g., Koss et al. 2011a), combined with dust obscuration of the central, AGN-dominated line-emitting regions (e.g., Koss et al. 2010).

The ultimate goal of the BAT AGN Spectroscopic Survey (BASS) is to provide the largest available spectroscopic sample of Swift BAT ultrahard X-ray (14–195 keV) detected AGNs. BASS DR1 (Koss et al. 2017) used mostly archival optical spectra for 641 AGNs from the 70-month BAT catalog (Baumgartner et al. 2013). This was complemented by detailed 0.5–200 keV spectral measurements using Swift, Chandra, and XMM-Newton for 838 AGNs (Ricci et al. 2017a). A further 102 near-IR (NIR; 1–2.5 μm) spectra of BAT AGNs were reported on by Lamperti et al. (2017). Other data include high spatial resolution NIR adaptive optics (AO) imaging (Koss et al. 2018); extensive continuum modeling of the MIR and far-IR (FIR) emission from WISE, IRAS, Akari, and Herschel (Ichikawa et al. 2017, 2019); radio emission in several resolution regimes (Baek et al. 2019; Smith et al. 2020); and host-scale molecular gas measurements (Koss et al. 2021). From these data, numerous links and correlations were investigated, such as between X-ray emission and high-ionization optical lines (Berney et al. 2015), and/or ionized gas outflows (Rojas et al. 2020). A major result was the realization that the Eddington ratio is a key parameter in some of these links and scaling relations (e.g., Ricci et al. 2017b; Oh et al. 2017; Ricci et al. 2018). Other BASS DR1 studies focused on AGN clustering (Powell et al. 2018); the surprisingly weak (or indeed insignificant) correlation between the X-ray photon index, Γ, and Eddington ratio (Trakhtenbrot et al. 2017); the most luminous obscured AGNs (Bar et al. 2017); optically unobscured AGNs with massive X-ray absorbing columns (Shimizu et al. 2018); BAT blazars (Paliya et al. 2019); and a search for BH binaries (Liu et al. 2020). Since making BASS DR1 public, the BASS team has pursued additional optical and NIR spectroscopy toward the second data release of BASS (DR2), to obtain higher spectral resolution (R ≳ 1500) over the broadest possible spectral range (i.e., 3200–10000 Å), for the most complete sample of AGNs drawn from the 70-month BAT catalog. Below we present an overview of the BASS DR2, including a summary of our survey design and goals, data sets, key scientific results, and a comparison to other large AGN surveys. A more detailed, technical description of the DR2 observations and data is provided in Koss et al. (2022a). Throughout this work we adopt Ω_M = 0.3, Ω_Λ = 0.7, and H₀ = 70 km s⁻¹ Mpc⁻¹. To determine the extinction due to Milky Way foreground dust, we use the maps of Schlegel et al. (1998) and the extinction law derived by Cardelli et al. (1989).

BASS DR2 provides optical spectra and associated redshifts, broad and narrow emission line measurements, velocity dispersions, and derived quantities—particularly BH mass and Eddington ratio estimates—for a nearly complete (e.g., >95%) sample of the 858 AGN from the Swift BAT 70-month survey.

2.1. Major Science Goals and Key Measurements

2.2. Revised 70-month AGN Catalog

We briefly review the AGN catalog changes compared to BASS DR1, with further details provided in Koss et al. (2022a). The published 70-month BAT catalog (Baumgartner et al. 2013) is composed of 1210 ultrahard X-ray sources, including 822 classified as AGNs or associated with a galaxy and likely an AGN, 287 Galactic sources (e.g., high-mass X-ray binary, low-mass X-ray binary, cataclysmic variable, pulsar), 19 clusters, and 82 unknown sources. In the BASS DR1 (Ricci et al. 2017a, see Appendix A) new AGN candidates were identified among BAT detections based on WISE and soft X-ray data to increase the number of 70-month AGNs to 838.

However, even after the DR1, 44 unknown BAT sources, typically near the Galactic plane (∣b∣ < 10°), had not been associated with counterparts. Of these sources, 22 were found to be AGNs. Fifteen of these unknown sources were subsequently identified as Galactic with NuSTAR and/or Chandra follow-up surveys (e.g., Yukita et al. 2017; Kennedy et al. 2020; Halpern et al. 2018) and/or based on optical spectroscopy obtained in the BASS/DR2 (see Koss et al. 2022a, for details). Another two sources that were classified as AGNs in the DR1 based solely on their hard X-ray spectra were found to be Galactic.

Unfortunately, their were still seven sources that lie very close to the Galactic plane (∣b∣ < 3°) owing to their very high extinction values (A_V ∼ 5 − 43 mag) and very high source confusion in the optical/NIR (i.e., multiple stars per 1 arcsec² area), which made optical and NIR spectroscopy of the counterpart impractical. The number of AGNs is 858 after these updates since two sources in the DR1 were discovered to be Galactic. ³⁷

For consistency with BASS DR1 and earlier studies, we classify AGNs according to the presence (or absence) of broad emission lines based on optical spectroscopy. Specifically, Sy1 are AGNs with broad Hβ line emission, Sy1.9 have narrow Hβ and broad Hα, while Sy2 AGNs have both narrow Hβ and narrow Hα (including small numbers of LINERs and AGNs in H₂-dominated regions). For beamed AGNs, the types include those with the presence of broad lines (BZQ), only host galaxy features lacking broad lines (BZG), or traditional continuum-dominated blazars with no emission lines or host galaxy features (BZB).

In addition to unbeamed AGNs, the Swift BAT survey includes also beamed and/or lensed AGNs, which are important to separate for most scientific analyses. The DR2 has 105 beamed AGNs as identified by their multiwavelength emission, and particularly radio emission, and/or Gamma-ray emission detected by Fermi (e.g., Paliya et al. 2019). This includes both blazars, where the boosted continuum emission completely dominates the (rest-frame) UV/optical regime and no significant emission lines are seen, and flat-spectrum radio quasars (FSRQs), which do show broad emission lines (e.g., Paliya et al. 2019). There are additionally two lensed AGNs. One of the beamed AGNs (SWIFT J1833.7−2105, aka PKS 1830−211 at z = 2.5) is also lensed (Lidman et al. 1999)) by a foreground galaxy. Thus, the total sample of 858 includes 752 unbeamed AGNs, 104 beamed AGNs, 1 beamed and lensed AGN, and 1 lensed and unbeamed AGN.

An X-ray luminosity and redshift plot of the BASS DR2, with the newly revised redshifts and AGN classifications, is shown in Figure 1. The figure also shows a few other deep and narrow X-ray AGN surveys. We discuss these samples and their comparison in more detail in Section 5.3; however, it is clearly evident that BASS provides a natural low-redshift benchmark for distant surveys.

Zoom In Zoom Out Reset image size

2.3. Survey Strategy and Observations

The sky distribution of all BASS DR2 optical spectroscopic observations is presented in Figure 2. Table 1 summarizes all of the BASS DR2 data, with further details provided in Koss et al. (2022a). The majority of the spectra used for the catalog measurements presented in DR2 papers come from either the Double Beam Spectrograph (DBSP) mounted on the Palomar Hale 5 m telescope (402 AGNs, mainly in the northern hemisphere) or the X-shooter spectrograph at the Very Large Telescope (VLT; 211 sources, mainly southern). In terms of DR2 spectroscopic targeting, our goals were to (1) provide the largest possible sample of BH mass measurements from either broad Balmer emission lines or stellar velocity dispersion measurements and (2) cover the broadest possible spectral range (e.g., 3200–10000 Å) for emission-line measurements, for the entire catalog of 858 AGNs. In practice, unless one uses echelle instruments (e.g., X-shooter), the latter goal requires either spectra with broad wavelength coverage and lower resolution or, alternatively, multiple, higher-resolution gratings and narrower wavelength coverage. The former goal motivates a spectral resolution of R ≳ 1000 (Δv ≲ 300 km s⁻¹ for broad lines), and yet higher resolution (>2000) for stellar velocity dispersion measurements. Our targeting strategy was further complicated by the fact that for a significant number of targets we did not know a priori the broad/narrow-line nature of the sources, and thus whether high-resolution stellar velocity dispersion measurements were required. Repeated observations were therefore done if either the S/N of the broad Balmer lines (Hβ or Hα) or the S/N and/or spectral resolution of the stellar absorption features were too low. Repeated observations with higher spectral resolution but limited spectral range were primarily done for obscured AGNs (Sy1.9 and Sy2) to measure velocity dispersions (and deduce BH masses), as velocity dispersion measurements are much more difficult and less reliable for AGN-dominated continuum. We did not reobserve targets with acceptable spectra and measurements from the Sloan Digital Sky Survey (SDSS; York et al. 2000) included in Data Release 16 (DR16; Ahumada et al. 2020).

Table 1. Summary of BASS DR2 Data

Telescope	Instrument	Total	Range (Å)	Slit Width (arcsec)	R
(1)	(2)	(3)	(4)	(5)	(6)
Palomar	DBSP	502	3150–10500	1.5	1220/1730
VLT	X-shooter	233	2990–10200	1.6/1.5	3850/6000
APO	SDSS	177	3830–9180	3	1760/2490
du Pont	BC	119	3000–9070	1	480
Archival a	Various	90	Various	Various	Various
VLT	FORS2	70	3400–6100	1	830
SOAR	Goodman	67	4560–8690	1.2	890/1630
Keck	LRIS	21	3200–10280	1	1280/1810
Magellan	MAGE	12	3300–10010	1	3850
VLT	MUSE	6	4800–9300	2	1850/3150
Velocity Dispersion Setups b
SOAR	Goodman	86	7900–9070	1.2	4720
Palomar	DBSP	66	3970–5499/8050–9600	2	2170/4720
NIR DR2
VLT	X-shooter	168	10240–24800	0.9	5400
Magellan	FIRE	65	8000–25000	0.6	6000

Notes. Column (1): telescope. Column (2): instrument. Column (3): total number of DR2 spectra observed with this setup. Column (4): wavelength range for the most common setup with the telescope. Columns (5)–(6): slit width and resolving power for the most common setup. In some cases larger or smaller slit widths (e.g., 15 versus 2'') were used, resulting in different resolutions. See Koss et al. (2022a) for a detailed list of instrument setups. Two values are listed when the instrument had both a blue and red arm with different settings. Resolving power is wavelength dependent in some cases, and so the values are given at 5000 and 8500 Å depending on the spectral range.

^a The archival sample is from earlier surveys that were not included in the DR1, including ROSAT AGNs that overlap with BASS in unpublished or published (Grupe et al. 2004) works, from the Palermo surveys of Swift BAT AGNs (Rojas et al. 2017), or as part of an atlas of low-redshift AGNs (Ho & Kim 2009). While not typically used in catalog measurements because of new DR2 data, we include these spectra for long-term studies of changing-look AGNs (e.g., MacLeod et al. 2019). ^b These setups were done for velocity dispersion measurements of obscured AGNs (e.g., Sy1.9 and Sy2).

Download table as:  ASCIITypeset image

Zoom In Zoom Out Reset image size

As the goal of BASS is to provide the best and most complete set of derived measurements (e.g., for narrow lines, broad lines, velocity dispersions), we did not require that all studies use the same single best spectra for an AGN in the DR2 as was done in the DR1. So for instance, a single Sy1.9 AGN may have a broad Hα measurement from SOAR/Goodman using the lower-resolution 400 line mm⁻¹ setting, along with measurements of narrow emission lines, but have a velocity dispersion measurement from CaT using the 1200 line mm⁻¹ setting. We also did not specifically exclude sources with high Galactic extinction if there was an obvious optical counterpart. This resulted in some spectra with very high extinctions (between A_V = 5 and 10 mag) that are only suitable for basic emission-line identification/classification and for redshift measurement.

2.4. Comparison with BASS DR1

2.5. NIR DR2

2.6. Survey Uniformity and Completeness

The DR2 papers listed in Table 2 provide a combination of large data sets of measurements and derived quantities, as well as scientific results for either the entire DR2 sample or considerable subsets of data. We provide a short review of these papers here, to explain the different measurements and catalogs provided in the various papers and to highlight key scientific findings. In some papers (e.g., narrow emission line measurements) a very wide spectral range was critical (3200–10000 Å), while in others (e.g., velocity dispersions) high spectral resolutions were critical; therefore, sometimes spectra with high spectral resolution but narrow wavelength range were used (e.g., for the calcium triplet, CaT, 8498, 8542, and 8662 Å). All DR2 spectra will be provided at the BASS website. ³⁹

The main, detailed DR2 catalog paper (BASS XXII, Koss et al. 2022a) gives an updated list of counterparts and a complete summary of observing characteristics and reduction procedures for the 1449 optical spectra of the 858 AGNs from the BAT 70-month sample (Table 1). It discusses various issues for basic AGN observational studies (source confusion, chance alignment of multiple AGNs leading to flux boosting, dual AGNs, etc.), including a revised list of beamed and lensed AGN identifications. This paper also provides a set of overall best-derived measurements (e.g., redshifts and distances, bolometric luminosities, and M_BH) since multiple BH mass estimates are available for some sources from various tracers (i.e., different broad lines, velocity dispersions, and/or high-quality literature measurements). The overall best available measurements from this catalog are then used consistently in subsequent scientific papers. Narrow optical emission line measurements over the range 3200–10000 Å (e.g., $\left[\mathrm{Ne}\,{\rm\small{V}}\right]\,\lambda 3426$, Hβ, $\left[{\rm{O}}\,{\rm\small{III}}\right]\,\lambda 5007$, Hα, $\left[{\rm{N}}\,{\rm\small{II}}\right]\,\lambda 6583$, $\left[{\rm{S}}\,{\rm\small{III}}\right]\,\lambda 9531$) are presented in BASS XXIV (Oh et al. 2022). In that paper, we employ a full-range spectral fitting procedure, incorporating both stellar population synthesis models and empirical stellar templates, which deblends complex nebular emission-line features from the stellar components. We also study AGN subtypes as a function of X-ray column density; strong-line ratio diagnostic diagrams (BPT) and their links to Eddington ratio, and line width comparisons between X-ray BAT AGNs and optical SDSS AGNs. Broad emission line measurements (Hα, Hβ, Mg ii λ2798, and C iv λ1549) and derived quantities are the focus of BASS XXV (Mejía-Restrepo et al. 2022). This paper includes virial estimates of BH mass (M_BH), which are used as the best BH mass measurement throughout DR2 and also allow estimates of the Eddington ratio (L_bol/L_Edd). The use of Mg ii λ2798 and C iv λ1549 is reserved for the beamed AGNs with discernible broad lines, which are at higher redshifts (e.g., z ≳ 1), where rest-frame Hβ falls outside of the optical range. This paper concludes that the innermost part of the broad-line region (BLR), which contributes the highest velocity emission, is preferentially absorbed in obscured AGNs ($\mathrm{log}({N}_{{\rm{H}}}/{\mathrm{cm}}^{-2})\gt 22$) and/or Sy1.9. This leads to a significant underestimation of the line flux compared to unobscured sources, which then strongly underestimates BH mass. These discrepancies typically exceed 1 dex and may reach 2 dex for heavily obscured AGNs ($\mathrm{log}({N}_{{\rm{H}}}/{\mathrm{cm}}^{-2})\geqslant 24$). We provide some prescriptions for corrections.

Central velocity dispersion measurements (e.g., from [Ca H + K ii λλ3935, 3968, Mg i, or CaT regions) of obscured systems (Sy1.9 and Sy2) are investigated in BASS XXVI (Koss et al. 2022b). This paper finds that BASS AGNs have much higher velocity dispersions than the more numerous optically selected narrow-line AGNs (i.e., ≈150 vs. ≈100 km s⁻¹), but also that BASS AGNs are not biased toward the highest velocity dispersions seen in massive ellipticals (i.e., >250 km s⁻¹). Additionally, despite sufficient spectral resolution to resolve the velocity dispersions associated with the bulges of relatively small SMBHs (∼10⁴−10⁵ M_⊙), we do not find a significant population of such AGNs, which (given the BAT flux limit) would have presented super-Eddington accretion rates.

Ananna et al. (2022) use the highly complete set of BASS DR2 measurements to derive the intrinsic XLF, BH mass function (BHMF), and Eddington ratio distribution function (ERDF), for both obscured and unobscured low-redshift AGNs using the BAT sample. It employs an elaborate forward-modeling approach to derive the intrinsic XLF, BHMF, and BHMF from the observed distributions, while accounting for various selection effects. We find that the intrinsic ERDF of narrow-line (type 2) AGNs is significantly skewed toward lower Eddington ratios than those of broad-line (type 1) AGNs, while the BHMFs of these subsamples are consistent with each other. This result offers insights into the geometric structure of the obscuring "torus" and lends support to the radiation-regulated unification scenario (Ricci et al. 2017b), which suggests that radiation pressure dictates the geometry of the dusty obscuring structure around an AGN. The XLF, BHMF, and ERDF are also used to investigate the AGN duty cycle in the low-redshift universe.

Pfeifle et al. (2022) investigate the relationship between MIR colors and X-ray column density. Heavily obscured BAT AGNs are found to be more heavily X-ray suppressed, displaying lower ratios of ${L}_{2\mbox{--}10\ \mathrm{keV}}^{\mathrm{obs}}$/L_{12 μm}, and they display "redder" MIR colors compared to unobscured AGNs. This paper develops diagnostic criteria that are designed to select both highly complete and highly reliable samples of heavily obscured AGNs ($\mathrm{log}({N}_{{\rm{H}}}/{\mathrm{cm}}^{-2})\gt 23.5$). We also derive expressions relating the luminosity ratios and column density, to predict the AGN column density in lower count-rate X-ray SEDs, where detailed spectral modeling is impossible. These diagnostics could be used on future samples of AGNs, such as those being discovered by eROSITA (Predehl et al. 2021), to efficiently distinguish between heavily obscured and unobscured AGNs.

Den Brok et al. (2022) provide a detailed analysis of the NIR coronal lines (CLs; ionization potential χ > 100 eV) and test their usage as indicators of AGN activity by comparing their strength, in particular that of $\left[\mathrm{Si}\,{\rm\small{VI}}\right]\,\lambda 19640$, to the X-ray flux. A key finding is that CLs correlate more tightly (i.e., smaller scatter) with the X-ray fluxes than with the optical $\left[{\rm{O}}\,{\rm\small{III}}\right]\,\lambda 5007$ line fluxes. Even in these bright AGNs, in only about half of the sources is a CL detected, limiting the extent to which CLs can be used as tracers of AGN activity. This study finds a clear trend of line blueshifts with increasing ionization potential in several CLs, such as $\left[\mathrm{Si}\,{\rm\small{VI}}\right]\,\lambda 19640$, $\left[\mathrm{Si}\,{\rm\small{X}}\right]\,\lambda 14300$, $\left[{\rm{S}}\,{\rm\small{VIII}}\right]\,\lambda 9915$, and $\left[{\rm{S}}\,{\rm\small{IX}}\right]\,\lambda 12520$, which elucidates the radial structure of the CL region.

Finally, Ricci et al. (2022) investigate the NIR BLR using Paα, Paβ, and He I λ10830 and associated virial BH mass estimates. The NIR regime is less affected by dust than the optical and can thus trace the innermost and fastest-moving BLR gas—even in the presence of mild obscuration. The study finds that the velocities of the BLR gas as estimated from the FWHMs of Hα and the NIR lines in Sy1–1.9 agree and are independent of the level of BLR extinction or obscuration (for $\mathrm{log}({N}_{{\rm{H}}}/{\mathrm{cm}}^{-2})\lt 23.5$), but the broad-line luminosities are suppressed with increasing obscuration, biasing virial-based M_BH estimates. The latter finding is in agreement with the conclusion of Mejía-Restrepo et al. (2022; see above). The line luminosity decrement and the obscuration level at which it occurs change as a function of wavelength, with Hα experiencing a higher decrement than Paα (above $\mathrm{log}({N}_{{\rm{H}}}/{\mathrm{cm}}^{-2})\simeq 21.0$ and 21.9, respectively). Thus, we caution against relying solely on Hα-based single-epoch BH mass estimates when $\mathrm{log}({N}_{{\rm{H}}}/{\mathrm{cm}}^{-2})\gtrsim 21$ and on NIR lines when $\mathrm{log}({N}_{{\rm{H}}}/{\mathrm{cm}}^{-2})\simeq 22$. A less biased proxy for the BLR radius in virial-based M_BH should be used at higher N_H, such as L_X.

Beyond the complete coverage with optical spectroscopy and the extensive NIR spectroscopy that are the main components of the BASS DR2, the BASS project aims to obtain and analyze large multiwavelength data sets for the BAT AGNs, in the X-ray, UV, optical, IR, FIR/submillimeter, and radio regimes. These include ongoing legacy BASS campaigns, past observations through BASS and the community, and various all-sky surveys (e.g., GALEX, GAIA, 2MASS, WISE, Akari, IRAS). The status of targeted observational programs for the 70-month AGNs as of 2021 July is summarized in Table 3 and on the BASS website. ⁴⁰

In addition to this, partial sky coverage exists from several wide-field surveys for hundreds of AGNs. Multiband high-quality optical imaging (<2'') exists for the majority (>80%) of BASS AGNs from the SDSS, the Panoramic Survey Telescope and Rapid Response System (Pan-STARRS; Chambers et al. 2016), the DESI Legacy Imaging Surveys (Dey et al. 2019), the VLT Survey Telescope ATLAS (Shanks et al. 2015), and targeted studies like Koss et al. (2011b). In the NIR, coverage exists in the VISTA Hemisphere Survey (VHS; McMahon et al. 2013) and the UKIRT Hemisphere Survey (UHS; Dye et al. 2018). In the radio, coverage at 1.4 GHz exists for more than half of the BAT AGNs (Wong et al. 2016) from the Faint Images of the Radio Sky at Twenty cm (FIRST; Becker et al. 1995) and NRAO VLA Sky Survey (NVSS; Condon et al. 1998).

The BAT AGNs have the largest coverage in the X-rays and UV, with all 858 observed with Swift/XRT and also observed with Swift/UVOT in the UV and optical. This results in a vast database of simultaneous X-ray and UV observations, obtained since the launch of Swift in 2004. For instance, for broad-line (Sy1) AGNs there are 32,184 distinct observations, due to the slewing nature of Swift and their intensive coverage in legacy observations of BAT AGNs.

As the brightest ultra-hard-X-ray-selected AGNs in the sky, many archival observations are available from several other X-ray telescopes. In particular, many BAT AGNs have also been observed as part of 20 ks filler observations in the NuSTAR BAT Legacy Survey, ⁴¹ which continues to observe approximately six BASS sources each month. A Chandra Cool Target program, ⁴² which started in 2019 January, is also observing nearby BAT AGNs (z < 0.1).

Another subset of observing programs focuses on high spatial resolution imaging (∼100 pc) that can be achieved for nearby BASS AGNs (z < 0.1). This includes a recent HST SNAP program with the Advanced Camera for Surveys (ACS), which has obtained i-band (F814W) images for 154 DR1 BAT AGNs at z < 0.1 (Kim et al. 2021). A large HST SNAP program, approved through 2022, aims to obtain near-UV (<3000 Å) imaging of BASS AGNs, followed up with simultaneous X-ray and UV/optical observations of the nuclear AGN emission with Swift and ground-based optical imaging in griz. Koss et al. (2018) published 98 Keck/NIRC2 AO-assisted NIR observations (in the Kp band) for a volume-limited sample of AGNs (z < 0.1), along with many archival HST NIR observations of more nearby galaxies available from earlier HST/NICMOS studies (e.g., Hunt & Malkan 2004). Finally, approved high-resolution AO NIR imaging and spectroscopy are also underway through 2022 using Keck/NIRC2, Keck/OSIRIS, and Gemini/GSAOI, with a focus on hidden galaxy mergers and dual, small-separation AGNs in obscured systems.

Finally, another set of survey programs are broadly focused on connecting the key star-formation-related properties of the AGN hosts, such as SFR and molecular and atomic gas, with AGN activity in nearby AGNs, using high-resolution and high-sensitivity observations over the IR−millimeter−radio spectral regimes. These programs include earlier studies carried out with Spitzer (e.g., Weaver et al. 2010) and Herschel (e.g., Mushotzky et al. 2014), which focused on nearby BAT AGNs and were then followed up in the submillimeter and radio. The Herschel program to measure star formation observed 317 of the nearest BAT AGNs (z < 0.05), which were later followed up with more recent measurements of host galaxy molecular gas using CO lines (Koss et al. 2021) and 22 GHz (e.g., Smith et al. 2020) observations with the JVLA. A yet-nearer-distance sample (z < 0.025) is being targeted for Hi mapping using the JVLA. More molecular gas observations using APEX have also been approved for these sources, through 2022. A program to obtain 100 pc resolution CO (2–1) measurements using ALMA was done for 33 nearby and luminous AGNs. High spatial resolution radio observations (0.2–0.5 pc resolution) that form a complete volume-limited sample out to 40 Mpc for AGNs above ${L}_{14-195\ \mathrm{keV}}^{\mathrm{obs}}$ > 10⁴² erg s⁻¹ has also been done for a sample of 37 objects using C-band Very Long Baseline Array (VLBA).

The BASS survey is spectroscopically complete for 100% (858/858) of the AGNs identified in the 70-month BAT all-sky survey outside of sources deep within the Galactic plane (∣b∣ < 3°), which we were unable to target. The BASS DR2 reports redshifts for 99.9% (857/858) of the AGNs, excluding only one continuum-dominated blazar with a foreground Galactic star. This includes 47 redshifts reported for the first time. Outside of the Galactic plane (∣b∣ > 10°), the survey completeness in BH mass measurements from broad lines or stellar velocity dispersion is 98% for all unbeamed AGNs because of the typically lower extinction in these regions. The remaining sources without BH mass measurements are mostly double-peaked and/or asymmetric broad-line AGNs and high-redshift Sy2 (z > 0.1), where high-quality velocity dispersion measurements are difficult. For beamed AGNs we report BH masses only for the BZQ class based on broad lines; for continuum-dominated blazars (BZB) velocity dispersion measurements are difficult due to the AGN-dominated continuum.

5.1. Survey Completeness and Measurables

In this section, we briefly illustrate the distributions of some of the key AGN-related properties of the BAT 70-month AGNs, as determined through the BASS DR2 measurements. We also briefly discuss the reliability and limitations of the measurements. A more elaborate forward-modeling approach accounting for various survey selection effects associated with a flux-limited survey with various levels of obscuration is described in Ananna et al. (2022).

Summaries of the typical bolometric luminosities, BH masses, Eddington ratios, and line-of-sight hydrogen column densities are provided in Tables 4 and 5 for unbeamed and beamed AGNs, respectively. Their BH mass versus Eddington ratio parameter space is shown in Figure 3. Overall, the completeness is quite similar for Sy1, Sy1.9, and Sy2 (∼98%) outside the Galactic plane.

Table 4. Summary of Unbeamed AGN Properties

Type	N	N	z	N ${}_{{M}_{\mathrm{BH}}}$	N ${}_{{M}_{\mathrm{BH}}}$	% Meas. MBH	% Meas.	log MBH	log Lbol	log Lbol/LEdd	log NH
		∣b∣ > 10°			∣b∣ > 10°		∣b∣ > 10°	(M⊙)	( erg s−1)		(cm−2 )
(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)
Sy1	359	318	0.050 ± 0.003	350	311	97	98	7.81 ± 0.04	44.87 ± 0.04	−1.17 ± 0.03	20.0 ± 0.05
Sy1.9	101	86	0.030 ± 0.004	97	84	96	98	7.98 ± 0.06	44.59 ± 0.08	−1.61 ± 0.09	22.28 ± 0.13
Sy2	292	259	0.029 ± 0.003	275	253	94	98	8.06 ± 0.04	44.50 ± 0.04	−1.71 ± 0.04	23.27 ± 0.05
Total	752	663	0.038 ± 0.002	722	648	96	98	7.96 ± 0.03	44.67 ± 0.03	−1.42 ± 0.03	21.98 ± 0.06

Note. Summary of the medians and standard error of the median for different populations of unbeamed AGNs. Column (1): AGN optical type based on the presence of broad Hβ and Hα. Column (2): total for the whole sample. Column (3): total excluding the Galactic plane region ∣b∣ < 10°, where high optical extinction makes measurements more difficult. Column (4): median redshift from optical lines. Columns (5)–(8): number of unique AGNs with M_BH measurements and excluding the Galactic plane region ∣b∣ < 10°, where high optical extinction makes measurements more difficult; also listed as percentages. Columns (9)–(12): median M_BH, L_bol, L_bol/L_Edd, and $\mathrm{log}({N}_{{\rm{H}}}/{\mathrm{cm}}^{-2})$ for the sample.

Download table as:  ASCIITypeset image

Table 5. Summary of Beamed AGN Properties

Type	N	N, ∣b∣ > 10°	z	N ${}_{{M}_{\mathrm{BH}}}$	% Meas. MBH	log MBH	log Lbol	log Lbol/LEdd	log NH
(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)
BZQ	74	63	0.88 ± 0.12	67	91	8.83 ± 0.09	47.66 ± 0.16	0.38 ± 0.12	20 ± 0.11
BZB	22	18	0.13 ± 0.02				45.81 ± 0.12		20.57 ± 0.12
BZG	8	6	0.07 ± 0.02				45.11 ± 0.20		20.81 ± 0.11
Sy1/lensed	1	1	0.65	1	100	8.79	47.18	0.21	20
BZQ/lensed	1	0	2.51				49.49		22.77
Total	106	88	0.33 ± 0.10	68		8.83 ± 0.09	46.53 ± 0.14	0.38 ± 0.12	20.54 ± 0.08

Note. Summary of the medians and standard error of the median for different populations of beamed and/or lensed AGNs. Column (1): AGN optical type based on presence of broad lines (BZQ), only host galaxy features lacking broad lines (BZG), or traditional continuum-dominated blazars with no emission lines (BZB), or lensing. Column (2): total for the whole sample. Column (3): total excluding the Galactic plane region ∣b∣ < 10°, where high optical extinction makes measurements more difficult. Column (4): median redshift from optical lines. Columns (5)–(6): number of unique AGNs with M_BH measurements and percentages. Columns (7)–(10): median M_BH, L_bol, L_bol/L_Edd, and $\mathrm{log}({N}_{{\rm{H}}}/{\mathrm{cm}}^{-2})$ for the sample.

Download table as:  ASCIITypeset image

Zoom In Zoom Out Reset image size

As a survey of the nearest and brightest hard-X-ray-selected AGNs in the sky, the measurement-associated uncertainties on BASS measurements are typically small. Apart from the beamed sources (e.g., blazars or "BZB") that lack emission lines and a handful of sources that are located in extremely crowded (Galactic plane) regions, the spectra of all BASS sources have multiple emission lines for robust redshift measurements. The uncertainties on BH mass determinations from σ_* are dominated by the systematics on the M_BH–σ_* scaling relation (e.g., ∼0.3–0.5 dex; Marsden et al. 2020) rather than on the scatter found in repeat observations (Koss et al. 2022b; ∼0.1–0.2 dex). Similarly, the uncertainties on BH masses derived through spectral analysis of broad lines may reach ∼0.3–0.4 dex (Peterson 2014), whereas the measurement uncertainty is much lower (i.e., 0.1 dex; Mejía-Restrepo et al. 2022). The uncertainties on N_H are ∼0.05 dex for $\mathrm{log}({N}_{{\rm{H}}}/{\mathrm{cm}}^{-2})$ < 23.5 and ∼0.3 for $\mathrm{log}({N}_{{\rm{H}}}/{\mathrm{cm}}^{-2})$ > 23.5 (Ricci et al. 2017a). For intrinsic X-ray luminosity, the errors are typically < 0.1 dex (Lanz et al. 2019), unless the AGNs are CT, in which case the typical errors can reach ≈ 0.4 dex (Ricci et al. 2015). The bolometric luminosities are calculated from the intrinsic luminosities in the 14–150 keV range as shown in Ricci et al. (2017a, see their Table 12), using a bolometric correction of 8 (see, e.g., Koss et al. 2022a). In this case, the uncertainties are in the range of 0.2 dex (e.g., Trakhtenbrot et al. 2017).

We also looked at relationships between the line-of-sight column density, as measured from X-ray data, and BH mass and Eddington ratio, as determined from optical spectroscopy (Figure 4). One clear takeaway is that unobscured (Sy1) AGNs occupy a region of higher Eddington ratios compared to obscured (Sy1.9 and Sy2) AGNs. This echoes the previous BASS DR1–based finding by Ricci et al. (2017b), which supports a scenario where radiation pressure is the main driver of the geometry of the (dusty) circumnuclear gas. There are no obvious trends between N_H and either M_BH or L_bol/L_Edd within each AGN optical subclass.

Zoom In Zoom Out Reset image size

5.2. Redshift Survey Biases

Looking more carefully into Figure 3, the unbeamed BASS AGNs typically occupy narrow regions in the BH mass—Eddington ratio plane, in different redshift intervals. These could be easily understood as a combination of survey (flux) limits, AGN physics, and demographics. For example, the most distant AGNs in our sample (z > 0.1) include almost no sources with small BHs (M_BH ≲ 10⁷ M_⊙) because these would have to be super-Eddington to be detected. ⁴³ Likewise, the lowest Eddington ratio sources (L_bol/L_Edd ≲ 10⁻³) are also almost exclusively at the highest-mass and lowest-redshift (z < 0.01) systems. However, while our sample is affected by strong low-mass and low accretion rate biases, there are no biases against super-Eddington non-beamed AGNs. Notably, such sources are extremely rare in the survey, suggesting that the Eddington limit remains meaningful despite the simplifications in its derivation. ⁴⁴

We show the Eddington ratio versus redshift in the top panel of Figure 5. Despite the aforementioned possible biases, Sy1 sources tend to have higher Eddington ratios (on average) than Sy1.9 or Sy2 sources, even when matched in redshift, though this difference decreases toward the highest redshifts in the sample (z > 0.05). The Sy1.9 and Sy2 classes follow the same distribution in L_bol/L_Edd rising sharply with redshift from L_bol/L_Edd ≈ 10⁻² at z = 0.01 to L_bol/L_Edd ≈ 6 × 10⁻². Conversely, the median Eddington ratio of Sy1 galaxies is nearly flat with redshift (L_bol/L_Edd ≈ (7 − 9) × 10⁻²).

Zoom In Zoom Out Reset image size

5.3. Comparison to Other Surveys

A comparison to other distant AGN X-ray surveys is shown in Figure 1. Due to its all-sky coverage, the BAT AGNs include the highest number of sources at z < 0.1, the most luminous unbeamed QSOs at 0.1 < z < 0.6, and a population of high-redshift beamed AGNs between z = 1 and 3.5. In luminosity space the survey overlaps well with medium-deep surveys (e.g., XMM XXL, Stripe 82, and COSMOS) at 0.5 < z < 3.

Among all-sky X-ray surveys, it is useful to compare the BAT AGNs to the past ROSAT survey in the soft X-ray band (0.1–2.4 keV; Truemper 1982). The more recent reprocessing of the second ROSAT all-sky survey (2RXS) source catalog (Boller et al. 2016) had a flux limit of ∼10⁻¹³ erg cm⁻² s⁻¹ at 0.1–2.4 keV, with ∼130,000 sources. Assuming Γ = 1.8, this flux limit corresponds to ∼2 × 10⁻¹³ over the BAT 14–195 keV band, which is ∼50× deeper than the 70-month BAT survey. Of these, 7005 ROSAT sources were cross-matched with the SDSS Data Release 5 (Anderson et al. 2007). Due to the soft X-ray sensitivity of ROSAT, the vast majority of these sources (6224/7005, or 89%) were broad-line AGNs, whereas the fraction within BASS for such (unbeamed) sources is 42%. In addition, the median redshift of the (subset of) ROSAT sources was z = 0.42, which is more than a factor of 10 more distant than the unbeamed BASS AGNs (z ≃ 0.037). The BASS overlap with ROSAT is 95% for Sy1 sources, down to 53% for Sy2, and only 30% for LINERs (see, e.g., Oh et al. 2022, for further details). A more comprehensive comparison between the 2RXS and the BAT AGNs is also available in Oh et al. (2018).

The concurrent eROSITA mission and its all-sky survey (0.2–8 keV; Predehl et al. 2021) are expected to eventually yield a few million AGNs and be roughly a factor of 100 times deeper (∼10⁻¹⁵ erg cm⁻² s⁻¹) than ROSAT, which means a particularly larger number of higher-redshift sources. The eROSITA Final Equatorial Depth Survey (eFEDS), with a depth of ∼ 10⁻¹⁴ erg cm⁻² s⁻¹ at 0.5–2 keV (Wolf et al. 2021) over 140 deg², provides some early insight into what one could expect for the AGN population to be surveyed by eROSITA. Specifically, 90% are unobscured ($\mathrm{log}({N}_{{\rm{H}}}/{\mathrm{cm}}^{-2})$ < 21.5), and the redshift distribution peaks at around z ≃ 1 (Liu et al. 2022), though redshift determination for the majority of sources is problematic until larger spectroscopic surveys are completed (e.g., via SDSS-V, Kollmeier et al. 2017; and/or VISTA/4MOST, Salvato et al. 2022). Thus, we expect that BASS will provide a bright complement of well-understood luminous nearby sources (z ≃ 0.037) that is less biased with regard to obscuration, but also missing the numerous distant AGNs to be detected by eROSITA.

To put the BASS sample in perspective, we also compare it to several other optical surveys of nearby luminous AGNs, including the SDSS quasars (Shen et al. 2011), SDSS Sy2 AGNs (Greene & Ho 2005), PG quasars (Boroson & Green 1992), and type 2 quasars selected using $\left[{\rm{O}}\,{\rm\small{III}}\right]\,\lambda 5007$ emission (Kong & Ho 2018). Compared to these samples, the BAT-selected AGNs are typically found at lower redshifts (z < 0.1). The Eddington ratios of broad-line (Sy1) AGNs are above the SDSS Sy2 AGNs, consistent with those of SDSS and PG quasars, but below the SDSS type 2 quasars. Among Sy1.9 and Sy2 types, only the highest-redshift BAT sources (z > 0.08) have similar Eddington ratios to the SDSS quasars and PG quasars. However, the SDSS type 2 quasars tend to have significantly higher Eddington ratios than BAT Sy1.9 or Sy2 types, despite overlap in redshift at z = 0.1. Finally, BASS Sy2 galaxies tend to have higher Eddington ratios than SDSS Sy2 galaxies.

The most luminous quasars in our sample are generally not found in other quasar samples. We investigated whether the most luminous BAT AGNs were selected by the SDSS quasars and type 2 quasar samples and found virtually no BAT AGNs in these samples. We focused specifically on the range of z < 0.3 and L_bol > 10⁴⁶ erg s⁻¹, which includes 18 unbeamed BAT AGNs, of which 8 are found within the SDSS footprint. Of these, only one source (SWIFT J1547.5+2050, aka 3C 323.1) is selected by the SDSS quasar sample. The other seven AGNs were not targeted for SDSS spectroscopy. In five cases the quasar is classified as a star in terms of colors, and in two cases the obscured (Sy2) AGNs are classified as galaxies, but no spectra were taken.

On the other hand, there are 14 SDSS quasars with L_bol > 10⁴⁶ erg s⁻¹, which are not detected by the BAT survey. The BAT detection limit at z = 0.3 for 90% sky coverage is equivalent to L_bol = 1.1 × 10⁴⁶ erg s⁻¹, assuming a simple conversion of L_bol = 8 × ${L}_{14-195\ \mathrm{keV}}^{\mathrm{obs}}$ erg s⁻¹ (e.g., Koss et al. 2017). Hence, all 14 SDSS quasars should be detected. It is possible that these undetected quasars may be part of a class of X-ray-weak quasars that have been found by several campaigns (e.g., Laor et al. 1997; Pu et al. 2020). Alternatively, the single, constant X-ray bolometric correction, rather than a luminosity-dependent one (e.g., Duras et al. 2020), may be too low for our sample. Techniques can be used to study known sources at ∼3 × fainter limits within BAT and would be ideal for this population (e.g., Koss et al. 2013). We reserve further discussion for future detailed studies. For the SDSS type 2 quasars, there is no overlap between the samples, with 26 AGNs in the type 2 quasars having L_bol > 10⁴⁶ erg s⁻¹ and 0/26 detected by BAT, when all should be detected.

We further compare the bolometric luminosities and BH masses of the BASS DR2 sample. Only the most luminous quartile of Sy1.9 and Sy2 AGNs reaches the average luminosities of the SDSS quasars at similar redshifts (z < 0.3). By comparison, roughly half of the BASS Sy1 sample occupies similar distributions in bolometric luminosity and BH mass to those of the SDSS quasars. The Sy1 galaxies have similar luminosities to the PG quasars, other than the least luminous quartile, but somewhat larger BH masses. The SDSS type 2 quasars occupy a region significantly above the BAT Sy1.9 and Sy2 galaxies in L_bol, reached only by the most luminous quartile of the Sy1 galaxies.

Finally, the BASS sample has a significant number of low-mass BHs (M_BH < 10⁷ M_⊙) that are not present in any of the other comparison samples. This feature of BASS is due to the higher spectral resolution (R > 2500) in the optical spectroscopy, which allows us to resolve spectral features tracing smaller BHs; the purely AGN-dominated selection in the >14 keV band, which allows us to study low-M_BH AGNs whose optical emission is host dominated; and the ability to study even the nearest AGNs (i.e., z < 0.01) that are saturated in SDSS imaging and thus excluded from the spectroscopic follow-up. A small number of nearby (z < 0.01) SMBHs have Eddington ratios as low as ≃10⁻⁵ (e.g., M81; Devereux 2019) and offer an opportunity to study the emission properties of advection-dominated accretion flows (Yuan & Narayan 2014). These radiatively inefficient accretion flows result in broadband SEDs that are markedly different from those characterizing standard, thin-disk accretion (see, e.g., Ryan & MacFadyen 2017).

We finally compare the BASS DR2 sample to MIR-selected AGNs. The standard WISE color cut (W1 − W2 > 0.8; Stern et al. 2012) identifies only 56% of the BAT AGN sample (482/858). The fraction of detections is highly dependent on AGN luminosity, with a much higher fraction of luminous AGNs detected (Figure 6; see also Ichikawa et al. 2017).

Zoom In Zoom Out Reset image size

We also compare the number of BASS AGNs in the WISE-selected AGNs drawn from the 30,093 deg² of extragalactic sky in the AllWISE Data Release (Assef et al. 2018). The AllWISE AGN study by Assef et al. (2018) provides an AGN catalog with 90% reliability (the "R90" catalog), selected purely using the WISE W1 and W2 bands, but with lower completeness. However, many of the BAT AGNs are excluded by default because they reside in galaxies that are extended in the Two Micron All Sky Survey (2MASS), a criterion adopted by the WISE AGN catalog to avoid contamination by separate, resolved parts of nearby galaxies. We find that only about 74/858 BASS DR2 sources overlap with the WISE R90 AGN catalog (9%), including 37 broad-line AGNs (Sy1–1.8), 6 narrow-line AGNs (Sy1.9–2), 28 beamed broad-line AGNs, and 3 continuum-dominated blazars (BZB). This corresponds to a WISE AGN detection fraction of 11.7% (37/314) for broad-line AGNs and only 1.5% for obscured BAT AGNs (6/393), though the detection fraction would be higher if extended galaxies were included in the WISE AGN catalog. We note again that the WISE detection fraction is strongly dependent on AGN luminosity, with no BAT AGNs with L_bol < 5 ×10⁴⁴ erg s⁻¹ found in the WISE AGN catalog.

When comparing the L_bol − z distribution of BAT- and WISE-selected AGNs in Figure 6, it appears that the WISE AGNs tend toward high redshifts at similar AGN luminosities, due to the requirement that they are point-like in 2MASS. If we further restrict the WISE AGNs to L_bol > 10⁴⁶ erg s⁻¹ and z < 0.3, there are six WISE AGNs above this limit from Barrows et al. (2021), five of which overlap with the BAT sample. On the other hand, we find that 6/18 (33%) of the unbeamed BAT quasars with z < 0.3 and L_bol > 10⁴⁶ erg s⁻¹ are in the WISE all-sky AGN catalog.

In summary, BAT is finding a broad range of nearby AGNs in terms of bolometric luminosity, BH mass, Eddington ratio, and particularly obscuration, including a significant population of low-L_bol/L_Edd, low-M_BH sources, with a well-characterized selection function and >99% complete spectroscopic coverage, making it a unique legacy sample for future AGN studies. Given the unique selection criteria provided by BAT and the complete, adaptive optical spectroscopy, it complements other legacy samples of nearby AGNs.

We present here an overview of the BASS DR2, with 1449 optical spectra, of which 1181 are released for the first time, for the 858 ultra-hard-X-ray-selected AGNs in the Swift BAT 70-month sample. In this special issue, we provide several immediate top-level scientific results and catalogs, including the following:

• 1.  
  A largely statistically complete sample with 99.6% and 98% of the brightest 858 ultrahard X-ray (14–195 keV) selected AGNs outside the Galactic plane having measured spectroscopic redshifts and BH mass estimates (respectively). The BH masses are derived from broad emission line (Mejía-Restrepo et al. 2022) or from stellar velocity dispersion measurements (Koss et al. 2022b).
• 2.  
  The 858 AGNs represent a uniquely complete census of nearby AGNs (z < 0.3), spanning 5–7 orders of magnitude in AGN bolometric luminosity (L_bol ∼ 10⁴⁰–10⁴⁷ erg s⁻¹), BH mass (M_BH ∼ 10⁵–10¹⁰ M_⊙), Eddington ratio (L_bol/L_Edd ∼ 10⁻⁵ − 10), and obscuration (N_H ∼ 10²⁰–10²⁵ cm⁻²). These AGNs are largely distinct from those found by other surveys, specifically with very little overlap even among nearby SDSS quasars or WISE AGNs.
• 3.  
  A large catalog of emission-line measurements from 3200 to 10000 Å (Oh et al. 2022) for the 858 AGNs and an additional 233 NIR spectroscopic measurements (1 − 2 μm; den Brok et al. 2022; Ricci et al. 2022).
• 4.  
  The first directly constrained BHMF and ERDF using both unobscured and obscured AGNs, in addition to a highly robust determination of the XLF (Ananna et al.2022).
• 5.  
  The significant bias toward underestimation of BH mass when using Hβ or Hα emission in obscured systems ($\mathrm{log}({N}_{{\rm{H}}}/{\mathrm{cm}}^{-2})\gt 21;$ see Mejía-Restrepo et al. 2022; Ricci et al. 2022).
• 6.  
  The ability of MIR emission to recover the X-ray column density (Pfeifle et al. 2022).

We hope that these initial results are only the beginning of the legacy value of the BASS project for understanding BH growth in nearby AGNs, and that the data products will be of lasting and general usefulness to the broader astronomical community. There are a variety of studies that can be done using this data set, such as focusing on SFRs, stellar masses, stellar population ages, dust reddening, metallicities, AGN-driven outflows, weak/faint emission lines, and links to morphological studies—all of which are not a significant part of the present data release. The broad wavelength coverage of the BASS sample is highly conducive to modeling SED with recent modeling tools (e.g., X-CIGALE; Yang et al. 2020). The future DR3 will focus in particular on fainter AGNs from the 105-month BAT catalog (Oh et al. 2018), which reaches flux limits 23% deeper than the 70-month catalog used for DR2, and for which follow-up observations are currently ongoing. We encourage the community to engage with the BASS data and team, to maximize the science output of this unique sample and data set.

We thank the reviewer for the constructive comments that helped us improve the quality of this paper.

BASS/DR2 was made possible through the coordinated efforts of a large team of astronomers, supported by various funding institutions, and using a variety of facilities.

We acknowledge support from NASA through ADAP award NNH16CT03C (M.K.); the Israel Science Foundation through grant No. 1849/19 (B.T.); the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation program, through grant agreement No. 950533 (B.T.); FONDECYT Regular 1190818 (E.T., F.E.B.) and 1200495 (E.T., F.E.B); ANID grants CATA-Basal AFB-170002 (E.T., F.E.B.), ACE210002 (E.T., F.E.B.), and FB210003 (C.R., E.T., F.E.B.); ANID Anillo ACT172033 and Millennium Nucleus NCN19_058 (E.T.); Millennium Science Initiative Program—ICN12_009 (F.E.B.); an ESO fellowship (M.H., J.M.); Fondecyt Iniciacion grant 11190831 (C.R.); the National Research Foundation of Korea grant NRF-2020R1C1C1005462 and the Japan Society for the Promotion of Science ID: 17321 (K.O.); Comunidad de Madrid through the Atracción de Talento Investigador grant 2018-T1/TIC-11035 (I.L.); support from the National Science Foundation under grant No. AST-1715512 (C.M.U.); YCAA Prize Postdoctoral Fellowship (M.B.); and from a Clay Fellowship administered by the Smithsonian Astrophysical Observatory and by the Black Hole Initiative at Harvard University, which is funded by grants from the John Templeton Foundation and the Gordon and Betty Moore Foundation (F.P.). This work was performed in part at the Aspen Center for Physics, which is supported by National Science Foundation grant PHY-1607611. We acknowledge the work done by the 50+ BASS scientists and Swift BAT team to make this project possible.

We acknowledge the various telescopes used in this paper. We are tremendously thankful to all the observing and support staff in all the observatories, and their headquarters, for their great assistance in planning and conducting the observations that made BASS/DR2 possible.

Specifically, BASS/DR2 is based on data obtained through the European Organisation for Astronomical Research in the Southern Hemisphere (ESO); the Palomar Observatory; the Southern Astrophysical Research (SOAR); the Kitt Peak National Observatory, managed by the US National Science Foundation's NOIRLab; the 6.5 m Magellan and the 2.5 m du Pont telescopes located at Las Campanas Observatory, Chile; the W. M. Keck Observatory at Maunakea, Hawaii; and the various stages of the Sloan Digital Sky Survey (SDSS).

A full list of observing facilities, program numbers, and their supporting bodies is provided in the main DR2 Catalog paper (Koss et al. 2022a; see the acknowledgments there).

A significant part of the BASS observations and work took place during the COVID-19 crisis. We thank the healthcare experts in communities around the world, for their tireless efforts to keep us all as safe and healthy as possible.

Facilities: Du Pont (Boller & Chivens spectrograph) - , Keck:I (LRIS) - , Magellan:Clay - , Hale (DBSP) - , NuSTAR - , Swift (XRT and BAT) - , VLT:Kueyen (X-shooter) - , VLT:Antu (FORS2) - , SOAR (Goodman). -

Software: Astropy (Collaboration et al. 2013), ESO Reflex (Freudling et al. 2013), IRAF (National Optical Astronomy Observatories 1999), Matplotlib (Hunter 2007), Numpy (van der Walt et al. 2011), Pandas (https://doi.org/10.5281/zenodo.3630805).