A Uniformly Selected, Southern-sky 6dF, Optical AGN Catalog

A complete census of active galactic nuclei (AGNs) is important for understanding the growth of supermassive black holes (SMBHs) and the co-evolution of SMBHs and their host galaxies via AGN feedback. While deep surveys have the highest number of AGNs, nearby, all-sky catalogs are important for understanding the low-redshift universe. AGN catalogs can be used to search for and study AGN-related phenomena, such as circumnuclear water maser emission (e.g., Zhu et al. 2011) and ultra-high-energy cosmic rays (UHECRs; e.g., Pierre Auger Collaboration et al. 2007, 2010). The interpretation of statistical correlation studies between the arrival direction of UHECRs and the sky locations of AGNs depends on the completeness, homogeneity, and purity of the AGN catalogs used. Furthermore, nearby AGNs can be studied in greater detail and newly identified AGNs can serve as additional targets for high-angular-resolution studies, especially for powerful telescopes in the Southern Hemisphere such as the Atacama Large Millimeter/submillimeter Array (ALMA; Kurz et al. 2002), Multi AperTure mid-Infrared SpectroScopic Experiment (MATISSE; Lopez et al. 2014), and GRAVITY (Gravity Collaboration et al. 2017).

Analysis of optical emission lines is one of the most robust methods of identifying AGNs, either via the presence of a broad Balmer line for Type 1 (unobscured) AGNs or via the ratios of the fluxes of the forbidden lines to those of Balmer lines for Type 2 (obscured) AGNs. Optical lines also provide measurements of AGN activity, since the [O iii] luminosity is a proxy for AGN bolometric luminosity, and of AGN obscuration, since the ratio of the Hα-to-Hβ line fluxes, called the Balmer decrement, increases with obscuration. Comparisons of the AGNs identified via different wavelengths is necessary to further assess the strengths and weaknesses of each method as well as a better understanding of the physical mechanisms responsible for the differences (e.g., Hickox & Alexander 2018). While optical observations of AGNs are highly reliable, due to their rich line information, they often miss low-luminosity and obscured AGNs. In comparison, infrared observations offer low obscuration bias but may miss star-formation-dominated AGNs and dust-poor AGNs. Similar to optical, X-ray observations are highly reliable and have low host contamination and low obscuration bias. However, X-ray observations can also miss low-luminosity and heavily obscured AGNs. If the AGN has no host contamination, and sufficient radio AGN emission, then the radio wavelength band should be ideal for obscuration bias-free analysis. However, only a small fraction of AGNs are radio-loud. AGNs identified with different methods should be combined to get a complete sample or as close as possible (e.g., Asmus et al. 2020). Our catalog can contribute to these studies, especially when compared with AGN catalogs from future X-ray telescopes, such as eROSITA, ³ Athena, ⁴ Lynx, ⁵ and the X-Ray Imaging and Spectroscopy Mission, formerly known as XARM, ⁶ and far-infrared (far-IR) radio surveys such as ALMA ⁷ and the Origins Space Telescope. ⁸

We have identified AGNs, using optical line ratios, from the Final Release of the Six-Degree Field (6dF) Galaxy Survey (6dFGS; Jones et al. 2004, 2009), a census of nearby galaxies in the Southern Hemisphere with decl. ≤ 0° and K_s ≤ 12.65. This work is an extension of the Zaw et al. (2019; ZCF, hereafter) all-sky, optical AGN catalog identified from the parent sample of Two Micron All-Sky Survey (2MASS) Redshift Survey (2MRS; Huchra et al. 2012) galaxy catalog, which contains 6dF galaxies with K_s ≤ 11.75. The Final Release of 6dFGS sample consists of 136,304 spectra (125,071 galaxies), more than 10 times larger than the subset of 11,762 galaxies included in 2MRS.

This paper is organized as follows. Section 2 describes our spectroscopic sample. Section 3 describes the methods we used to isolate the emission lines and identify AGNs. Section 4 describes our AGN catalog. We present the impact of the spectral signal-to-noise ratio (S/N) on AGN detection rates and statistically correct for the resulting incompleteness and inhomogeneity of AGN detection rates in Section 5. In Section 6 we summarize our catalog and findings.

We take 6dFGS (Jones et al. 2004, 2009) as the parent sample for constructing our southern sky optical AGN catalog. The survey targeted galaxies in the Southern Hemisphere (decl. ≤ 0°) with K_s ≤ 12.65 (median z = 0.053) are from the 2MASS (Skrutskie et al. 2006) catalog. The 6dFGS catalog consists of 136,304 publicly available digital spectra of 125,071 galaxies taken with the 6dF multifiber spectrograph, operated on the United Kingdom Schmidt Telescope (UKST). Figure 1 shows the sky distribution of the galaxies in the 6dFGS. The Galactic latitude cut of the 6dF data is ∣b∣ > 10°, as described by Jones et al. (2004), and the same is used for this work. The observations of 6dFGS were carried out at the UKST from 2001 to 2006 (Jones et al. 2004). The majority of the observations of the 6dFGS galaxies consist of two spectra, in the V band and the R band. Due to different gratings, data taken prior to 2002 October have a V band spanning 4000–5600 Å and an R band spanning 5500–8400 Å; later observations (∼80%) have a V band spanning 3900–5600 Å and an R band spanning 5300–7500 Å. V-band spectra have a resolution of 5–6 Å in FWHM, and R-band spectra have a resolution of 9–12 Å FWHM. The data do not have absolute flux calibration.

Zoom In Zoom Out Reset image size

In order to run the optical spectroscopic analysis for AGN candidates, we need spectra in both the V band, which covers the Hβ and [O iii] lines, and the R band, which covers the Hα and [N ii] lines. There are ∼3460 spectra with only one (either V or R) band available. Moreover, we find that there are ∼3400 spectra with negative mean S/N in either the V or R band. We remove the spectra that only have one band available and the ones with a negative mean S/N before we combine the V- and R-band data. To exclude the low S/N cases of the absorption lines, we made a conservative cut at a mean S/N ≥ 3 at both the V and R band, where the S/N is defined as the mean S/N over the full V band or R band. The data that passed the above criteria were then stitched to combine their V and R bands. We note that there are some galaxies with repeated observations—less than 10% of the total number of galaxies. A majority of these have two observations and a few have up to four observations. In case of repeated observations for the same galaxies (∼9000 galaxies), we keep the spectra with the highest S/N. After the above selection, we have 101,263 galaxies whose spectra passed the data quality control. ⁹

Note that a subset of the 6dF data (11,762 galaxies) were part of 2MRS (Huchra et al. 2012) and the AGNs in that sample are part of our all-sky optical AGN catalog (Zaw et al. 2019; ZCF). We do not repeat the data reduction and analysis of those here. The subset of 6dF data in 2MRS (hereafter 2MRS–6dF) and ZCF has a brighter magnitude cut at K_s ≤ 11.75 to match the uniform sky distribution based on the 2MRS survey. Therefore, the S/N of those spectra are higher than the full 6dF sample, with a magnitude cut of K_s ≤ 12.65. We show the redshift distribution in Figure 2, where the subset of 6dF from 2MRS is highlighted in dark gray. Redshifts (z) are from Jones et al. (2004, 2009). Due to the K_s requirement of 2MRS compared to 6dF, the objects of 2MRS–6dF are, in general, brighter than the rest of the 6dF objects. We therefore separate the 6dF sample into 2MRS–6dF and faint 6dF subsamples. We show the mean S/N distribution over all wavelengths in Figure 3. The 2MRS–6dF subsample has an overall higher S/N, with the mean S/N peaked at 20, while the full 6dF has significantly lower S/N, peaked at 10. Since the S/N values are wavelength dependent (e.g., Zaw et al. 2019), we show the mean and standard deviations of the S/N for the 6dF sample at various (continuum) wavelengths in Figure 4, where subsamples from 2MRS and the faint 6dF are presented to show the S/N differences.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

We use the wavelength range covered by both the V and R bands (5500–5600 Å or 5300–5600 Å) to merge the 6dF data that satisfy the S/N criteria into a single spectrum for each galaxy. Data are evenly resampled to log space to reduce the possible pixel shifts between different bands since there are some small wavelength shifts for the same spectral features in between the two arms due to the wavelength calibration procedure. We then use the arm with higher resolution, namely the V band, as a template to fit for the overlapped wavelength ranges in the R band using the penalized pixel-fitting (pPXF; Cappellari & Emsellem 2004) method to determine the number of pixels to be shifted and to merge the bands. Note that this is an automatic procedure, which stitches most of the data successfully. There are cases where the automatic stitching process fails due to poor data quality in one of the arms, or a large number of pixels in the overlap region whose values are zero, especially for pixels at the beginning of the R band and end of the V band. Examples of arm-merged 6dF spectra with different continuum S/Ns are shown in Figure 5. We identified ∼7000 spectra which fail automatic stitching by visual inspection. For the these cases, we had to merge the data manually, i.e., scaling the arms and shifting the pixels until no obvious discontinuity appeared between the two bands.

Zoom In Zoom Out Reset image size

We note that the spectra are not corrected for Earth's atmospheric absorption, i.e., the telluric absorption. The B band around 6860–6890 Å overlaps with the optical AGN identification emission lines, [N ii]–Hα complex, and can cause inaccurate line fluxes. This typically happens to the objects in the redshift range of 0.0407 < z < 0.05111. In those cases, we exclude these galaxies (∼14%) from our sample. However, when a clear broad Hα component is present, we still keep the galaxy as a potential AGN candidate although the fluxes of the broad Hα in these galaxies cannot be measured accurately due to the telluric contamination. In addition, when the galaxies are at higher redshifts, z ≥ 0.1428, the Hα line may shift out of the wavelength range, and we cannot assess whether or not they are AGNs.

The spectral fitting is performed on the bands-merged spectra in order to isolate the emission lines used in optical AGN identification. The spectral fitting procedure and validations are described in detail by Zaw et al. (2019). We briefly summarize it here.

3.1. Subtraction of the Host Galaxy Contribution to the Spectra

3.2. AGN Identification

After subtracting the continuum and absorption contributions from the host galaxy, we identify AGNs from the emission lines in the spectrum. Type 1 (broadline) AGNs are characterized by the presence of broadened hydrogen Balmer lines. As explained by the unified model (e.g., Antonucci 1993; Urry & Padovani 1995) this is the case when an AGN is unobscured, i.e., viewed face-on, and the Doppler broadened emission of the rapidly orbiting gas close to the supermassive black hole can be seen in the spectrum. If an AGN is seen closer to edge-on, the broadline regions are obscured by intervening gas and dust, leaving only narrow emission lines in the spectrum. In this case, the Type 2 (narrow line) AGNs can be identified from the flux ratios of the forbidden ([N ii] and [O iii]) lines to the Balmer (Hα and Hβ) lines.

We fit Gaussians to the emission lines in the spectrum and calculate the FWHM and the flux of each line. For the [N ii]-Hα complex where the lines are blended together, three or four Gaussians are used to fit for the emission components simultaneously. The corresponding Gaussian components are used to calculate the fluxes and FWHM of [N ii] and Hα lines. The details of the fitting procedure are described in Section 3.3 and are shown in Figure 6 of Zaw et al. (2019). Line fluxes are calculated under the Gaussian profiles in the wavelength bins within 3σ of the fitted peak, where σ is the fitted width of the Gaussian. Flux errors are calculated by adding in quadrature the value of each bin within the same wavelength range (i.e., 3σ) in the error spectrum.

Our method has been validated by comparing it with the results from Sloan Digital Sky Survey (SDSS) Data Release 8 (DR8; Aihara et al. 2011). The detailed description can be found in Zaw et al. (2019) and we briefly summarize the main procedures and results here. We first compared the fluxes and flux errors we derived for the SDSS galaxies in 2MRS with those from DR8 and found good agreement. We then compared our measurements for the galaxies that have spectra from both SDSS and 6dF. Since 6dF spectra do not have absolute flux calibration, we cannot compare the fluxes. However, we find that the flux ratios are consistent between SDSS and 6dF spectra for the same galaxies. Since 6dF spectra have lower S/N on average, requiring S/N ≥ 2.0 for the emission lines in 6dF spectra gives roughly the same AGN identification rates as requiring S/N ≥ 3.0 for the emission lines in SDSS spectra.

Because redder diagnostic lines such as [S ii] λλ6717, 6731 lines may shift out of the available spectral wavelength at high redshift (z ≥ 0.114), for completeness, we only use the bluer optical emission lines, i.e., broad and narrow Hα, narrow Hβ, [O iii] λ5007, and [N ii] λ6584, for AGN identification. We use the ratio between the flux and flux error, (S/N)_line, to evaluate the significance of the emission-line features. Following the detailed discussion in Zaw et al. (2019), where common galaxies with high S/N are available from SDSS, (S/N)_line ≥ 2.0 is used to further investigate the AGN candidates. An emission-line galaxy is defined as one having all four lines with (S/N)_line ≥ 2.0. When broad Hα has (S/N)_line < 2.0 while the other narrow lines all have (S/N)_line ≥ 2.0, a triple Gaussian is used to refit the [N ii]–Hα complex to derive the corresponding line fluxes and flux errors. Galaxies with significant ((S/N)_line ≥ 2.0) broad Hα components and having wide FWHM ≥ 1000 km s⁻¹ (e.g., Eun et al. 2017; Ho et al. 1997; Schneider et al. 2010; Stern & Laor 2012; Vanden Berk et al. 2006) are classified as Type 1 AGNs. As mentioned above, the [N ii]–Hα complex is contaminated by the telluric absorptions at certain redshifts. The telluric-contaminated galaxies can be potential Type 1 AGNs. When such a galaxy shows a wide enough and prominent (usually through visual identification) broad Hα we still identify it as a Type 1 AGN, although its broad Hα flux is underestimated. Normally Hβ lines are weaker than Hα lines, therefore we do not require broad components of Hβ lines as the indication of Type 1 AGNs. However, there are potential Type 1 AGNs with broad Hβ components but weaker broad Hα as well. Those (about a few hundred) are semi-manually identified by checking their FWHM, (S/N)_line, and visual inspection. As mentioned in Section 2, no flux calibration was performed on the data, therefore, the absolute flux values are not available in this work.

When no significant broad Hα or Hβ are present, we use the line flux ratio as first proposed by Baldwin et al. (1981), known as the Baldwin–Phillips–Telervich (BPT) diagram to distinguish the Type 2 (narrow line) AGN from the star-forming galaxies whose spectra also show emission features from [O iii], [N ii], Hα, and Hβ. Line ratios [N ii]/Hα and [O iii]/Hβ are used in the BPT diagram to identify the possible AGN candidates. We adopt the two demarcation lines, which are usually used to separate AGNs and star-forming galaxies, namely,

$$\begin{eqnarray}&&\mathrm{log}([{\rm{O}}\,{\rm\small{I}}\mathrm{II}]/{\rm{H}}\beta )\gt 0.61/(\mathrm{log}([{\rm{N}}\,{\rm\small{I}}{\rm{I}}]/{\rm{H}}\alpha )-0.47)+1.19\end{eqnarray} \tag{ 1 }$$

$$\begin{eqnarray}&&\mathrm{log}([{\rm{O}}\,{\rm\small{I}}\mathrm{II}]/{\rm{H}}\beta )\gt 0.61/(\mathrm{log}([{\rm{N}}\,{\rm\small{I}}{\rm{I}}]/{\rm{H}}\alpha )-0.05)+1.3.\end{eqnarray} \tag{ 2 }$$

The first is based on theoretical modeling of the maximal line ratios possible in star formation (Kewley et al. 2001), and the second is based on empirical studies of SDSS galaxies (Kauffmann et al. 2003). Galaxies with line ratios falling above the Kewley et al. (2001) line have forbidden line emission dominated by the AGN. Galaxies between the Kewley et al. (2001) and Kauffmann et al. (2003) lines are likely to be Type 2 AGN but can have contributions from both AGN and star formation galaxies. These galaxies are also referred as composite galaxies. We use the Kauffmann et al. (2003) line to identify AGNs for our catalog. If an AGN is also above the Kewley et al. (2001) line, we mark it as such.

As there are multiple AGN identification line ratios, such as [S II]/Hα and [O I]/Hα (Kewley et al. 2006) besides the main line ratio of [N II]/Hα, we explore the AGN identification consistency with different criteria for those galaxies whose [S II] and/or [O I] emission lines also have S/N ≥ 2.0. We show the results in Figures 6 and 7. In both plots, AGN identified from [N II]/Hα are marked as orange dots. The AGN identification using both [S II]/Hα and [O I]/Hα are in general agreement with the results of [N II]/Hα. As expected, some deviations occur from using different line ratios, with the shift of some galaxies being identified as AGNs to being identified as non-AGNs or vice versa. In general, only a small fraction of galaxies that were identified as AGNs by [N II]/Hα move into the non-AGN region when using [S II]/Hα and [O I]/Hα line ratios. A larger fraction of galaxies that are not AGNs by [N II]/Hα criteria are identified as AGNs in the [S II]/Hα and [O I]/Hα BPT diagrams. This indicates that [N II]/Hα provides a purer sample of AGNs. In addition, [N II] lines are stronger and have lower wavelengths, i.e., they fall within the 6dF wavelength ranges at higher redshifts than the [S II] and [O I] lines. Therefore, we use [N II]/Hα as the AGN identification line ratio for this catalog.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Our AGN catalog consists of 3116 Type 1 AGNs and 12,162 Type 2 AGNs, which satisfy the Kauffmann et al. (2003) criteria. A subsample of 3866 Type 2 AGNs also satisfy the Kewley et al. (2001) criteria. We summarize the AGN numbers and fractions of each subsample in Table 1. N_sample is defined as the total number of valid spectra of each subsample, which includes the spectra that are free from telluric contaminations and some strong Type 1 AGN spectra affected by the telluric absorption. As mentioned above, we made a cut on S/N ≥ 3 in the V band for Type 2 AGNs and S/N ≥ 3 in the R band for Type 1 AGNs. A few Type 1 AGNs were identified by visual inspection whose R-band spectra are with an S/N ≤ 3 but strong broad emission lines. We mark the objects in common with Zaw et al. (2019) in this catalog by adopting their 2MRS name, i.e., name of 2MASS ID (TMID). Galaxies without TMID values are identified AGN from the faint 6dF sample. The sky distribution of our AGN catalog is shown in Figure 8. We show the BPT diagram for narrow emission-line galaxies in Figure 9, where line flux S/N cut (S/N)_line ≥ 2.0 is applied to all four diagnostic lines. In Figure 10 we show the mean S/N of the AGN sample across the V band and R band as a function of redshift. Data quality quickly decreases with redshift, from a mean S/N of ∼30 at z = 0.0 to S/N of ∼13 at z = 0.06. AGNs with higher redshift z ≥ 0.06 have a mean S/N of ∼10.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

We use the Kewley et al. (2006) criteria to further delineate Seyferts and low-ionization nuclear emission-line regions (LINERs). Note that ∼2% of the AGN sample are missing their [S ii] lines at higher redshift, and the discussion about LINERs is limited to lower redshifts z ≤ 0.114 . We show the Seyferts and LINERs of 2MRS and the faint 6dF samples in Figure 11. The LINER-to-Seyfert ratio of 2MRS–6dF is higher than that from the faint 6dF subsample. This is likely because LINERs have weaker emission lines and they are preferentially lost in spectra with lower S/N (Zaw et al. 2019). As expected, the LINER-to-Seyfert ratio of 2MRS dropped from 1.01 to 0.56 in the faint 6dF sample, which has a much lower S/N. Note that we use the [O iii]/Hβ versus [N ii]/Hα BPT diagram to identify AGNs, so some emission-line galaxies that do not qualify as AGNs in the main diagram appear in the AGN region in the [O iii]/Hβ versus [S ii]/Hα diagram.

Zoom In Zoom Out Reset image size

A sample of the catalog is shown in Table 2. We provide the following properties of the AGNs.

• 1.  
  TARGETNAME: 6dF ID, a combination of R.A. and decl. in sexagesimal units;
• 2.  
  R.A.: right ascension in degrees;
• 3.  
  Decl.: decl. in degrees;
• 4.  
  z: redshift value from Jones et al. (2004, 2009).
• 5.  
  Type: AGN Type, T1 = Type 1, K01 = Type 2 satisfying Kewley et al. (2001) criteria, K03 = Type 2 satisfying Kauffmann et al. (2003) criteria;
• 6.  
  SN_V : S/N of the V-arm of the spectrum which contains the Hβ line. Since Hβ is the weakest of the four AGN identification lines, this is the measure of spectral quality that affects the ability to detect narrow emission lines for Type 2 AGN identification;
• 7.  
  SN_R : S/N of the R-arm of the spectrum which contains the Hα line, the measure of spectral quality that affects the ability to detect the broad Hα line for Type 1 AGN identification;
• 8.  
  TMID: 2MASS ID, a combination of R.A. and decl. in sexagesimal units, to mark the objects in common with Zaw et al. (2019); and
• 9.  
  SSP χ²: the reduced χ² from SSP fitting. We applied the same SSP χ² cut at 6.05 to keep 99% of the successful fits as in Zaw et al. (2019).

Table 2. The AGN Catalog (Illustrative Extract)

TARGETNAME	R.A. (deg)	Decl. (deg)	z	Type	${\mathrm{SN}}_{{\rm{V}}}$	${\mathrm{SN}}_{{\rm{R}}}$	TMID	SSP χ2
g0000113_050932	0.04709	−5.15876	0.019057	K03	12.6	9.6	00001131-0509313	1.26
g0000182_472923	0.07596	−47.48958	0.027322	K01	23.6	23.6	00001825-4729226	0.99
g0000261_490431	0.10858	−49.07539	0.067518	K01, Sy2	12.8	12.0		1.23
g0000356_014547	0.14849	−1.76318	0.024424	K03	19.9	17.9	00003564-0145472	1.56
g0001138_440043	0.30740	−44.01183	0.038964	K03	17.1	15.5	00011378-4400426	0.92
g0001361_144455	0.40020	−14.74867	0.037679	K01, LINER	21.5	28.8	00013605-1444548	1.14
g0001558_273738	0.48258	−27.62723	0.028330	K01	14.1	17.8	00015583-2737382	2.07
g0002039_332802	0.51617	−33.46728	0.028937	K03	33.4	32.6	00020386-3328023	1.37
g0002348_034239	0.64505	−3.71072	0.021498	Sy1	21.4	16.6	00023480-0342386	1.84
g0054323_404258	13.63458	−40.71603	0.024142	K03	26.7	27.3	00543231-4042578	0.62
g1211143_393327	182.80946	−39.55747	0.061064	Sy1	17.3	18.8	12111425-3933268	0.77
g1606162_291654	241.56742	−29.28167	0.015668	K03	68.1	71.9		4.72
g1901287_243406	285.36954	−24.56833	0.029157	Sy1	27.1	28.5	19012871-2434060	1.41
g1955064_392224	298.77538	−39.37258	0.019925	K01	52.7	52.3		4.06
g2124243_234134	321.10283	−23.69403	0.000307	Sy1	5.1	10.9		0.99
g2241589_373443	340.49258	−37.57853	0.028797	Sy1	23.3	24.5	22415822-3734426	0.79
g2312198_274019	348.08262	−27.67189	0.004918	K01	3.8	3.7		0.78
g2347461_372925#	356.94221	−37.49031	0.043774	Sy1	22.9	25.9		0.72
g1148026_184952*	177.01100	−18.83114	0.033478	Sy1	3.8	3.9		3.19
g1526535_310839t	231.72292	−31.14408	0.047590	Sy1	17.1	19.7	15265348-3108388	1.07

Note. TARGETNAME: 6dF ID, # indicates that local fitting was used to estimate the galaxy contribution; * indicates Type 1 AGNs identified from broad Hβ; t indicates the Type 1 AGNs may be contaminated by telluric absorption in the [N ii]–Hα complex; R.A.: R.A.; decl.: decl.; z: redshift; and Type: AGN type, Sy1 = Type 1, K01 = Type 2 with Kewley et al. (2001) criteria, K03 = Type 2 with Kauffmann et al. (2003) criteria. We use [S II] to classify the LINERs and Type 2 Seyferts (Sy2) by adopting Kewley et al. (2006) criteria. When [S II] falls within the 6dF wavelength range, a K01 galaxy that satisfies either the LINERs or Sy2 criteria is marked accordingly in the column "Type"; TMID: 2MASS ID, indicate common AGNs in Zaw et al. (2019); SSP χ²: reduced χ² from the SSP fitting. The full table also includes the emission-line widths, fluxes, and flux errors of all the AGNs in the catalog.

Only a portion of this table is shown here to demonstrate its form and content. A machine-readable version of the full table is available.

Download table as:  DataTypeset image

In the online version of our catalog, we provide extended information for the identified AGNs, namely the FWHM and flux of the broad Hα line, fluxes, and flux errors of the narrow lines used in AGN identification, and the (S/N)_line for each line. The additional information allows users to create their own customized AGN catalog using different AGN selection criteria. Note that the flux values are not absolute fluxes due to the un-flux-calibrated spectra. Therefore, the reported fluxes should only be used as ratios. AGNs with FWHM ≥ 4000 km s⁻¹ in the broad Hα line are likely to have underestimated FWHM values, as their broad lines are often lumpy and asymmetric which need multiple Gaussians to properly model the line profiles (e.g., Greene & Ho 2007).

Statistical correlation studies are an important use of AGN catalogs. To fully understand the results of such studies, the completeness and homogeneity of the catalogs needs to be assessed. Our analysis in Zaw et al. (2019) shows that the AGN detection rates are due to the overall continuum S/N of the spectrum. The line S/N is determined by the spectral quality (i.e., the continuum S/N) and the strength of the emission-line in a spectrum. The continuum S/N is a measure of the noise at a given wavelength and the ability to detect a line at that wavelength. Since data quality plays a crucial role in spectroscopically identifying AGNs, the AGN fraction is accordingly affected.

As mentioned above, the 6dF data consist of two bands, namely the V band and the R band. The Hβ and [O iii] lines are normally located in the V-band spectra and the Hα–[N ii] lines are normally located in the R-band spectra. As the redshift of full 6dF spans a larger range, it is more robust to monitor the mean (S/N) of each band rather than using a narrow range of spectra around the Hβ and Hα lines as we did in Zaw et al. (2019). The AGN detection rate as a function of continuum S/N is presented in Figure 12, where we also show the histograms of the S/N distributions. As seen in the left column, the AGN detection rates increase with S/N for both the full 6dF and the faint 6dF, until the rate flattens at a saturation value of S/N ∼ 20–50. The saturation S/N depends on the type of AGN. Type 1 AGNs have a saturation value of S/N ∼ 40, Type 2 AGNs that satisfy the Kewley et al. (2001) criteria have a saturation value of S/N ∼ 50, and Type 2 AGNs that satisfy the Kauffmann et al. (2003) criteria have a saturation value of S/N ∼ 20 in the faint 6dF. Type 2 AGN that satisfy the Kauffmann et al. (2003) criteria for the full 6dF sample follows a smooth increasing curve, which is dominated by the 2MRS galaxies, while a more complicated behavior is seen in the faint 6dF sample due to its lower S/N. As mentioned above, the full 6dF has an overall lower S/N compared to the 2MRS subsample; a threshold is observed for the AGN fraction, namely S/N ∼ 55. This behavior is consistent among the 2MRS–6dF, faint 6dF, and full 6dF sample. In the right column, we show the S/N distributions of the subsamples. As mentioned before, the 2MRS–6dF subsample has higher S/N on average.

Zoom In Zoom Out Reset image size

We show the AGN ratio across the sky (left panel) and its distribution as a function of redshift (right panel) in Figure 13. The inhomogeneity of the AGN detection rates is observed in both panels, which is strongly affected by the data quality. Data with lower redshifts (z < 0.07) show higher AGN rates due to their higher data quality (as shown in Figure 12). While data with higher redshifts tend to flatten out at an AGN detection rate around 0.15. The gap in redshift panel z shows unavailable data due to telluric absorption, the same effect is seen in Figure 10.

Zoom In Zoom Out Reset image size

In order to be used in a rigorous statistical analysis, e.g., for correlations with the arrival direction of cosmic rays, we need to address the detailed inhomogeneity and incompleteness due to the missing or telluric-contaminated spectra. We perform similar corrections as in Zaw et al. (2019). We fit the AGN fraction as a function of the continuum S/N using a linear correlation below the saturation S/N value, and a constant above, for Type 1 AGNs and Type 2 AGNs satisfying the Kewley et al. (2001) criteria. ¹¹ Here are the relations we obtain from the fits:

$$\begin{eqnarray}&&\begin{array}{l}{R}_{{fT}1}=(\mathrm{SN}1)\ast 0.00261-0.00128\ \mathrm{for}\,\mathrm{SN}1\lt 40\\ =\ 0.108\equiv {R}_{{sT}1}\qquad \mathrm{for}\,\mathrm{SN}1\geqslant 40,\end{array}\end{eqnarray} \tag{ 3 }$$

$$\begin{eqnarray}&&\begin{array}{l}{R}_{{fT}2}=(\mathrm{SN}2)\ast 0.00346-0.00184\ \mathrm{for}\,\mathrm{SN}2\lt 55\\ =\ 0.192\equiv {R}_{{sT}2}\qquad \mathrm{for}\,\mathrm{SN}2\geqslant 55,\end{array}\end{eqnarray} \tag{ 4 }$$

where SN1 (SN2) is the continuum S/N for the R-band arm (V-band arm), i.e., containing the Hα (Hβ) line; R_fT1 (R_fT2) is the AGN fraction for the given galaxy's S/N value for Type 1 (Type 2) AGNs; and R_sT1 (R_sT2) is the saturation value for the Type 1 (Type 2) AGN fraction. We assign a Type 1 AGN likelihood of zero for a non-AGN whose S/N (R-band arm) value is above 40. Similarly, we assign a Type 2 AGN likelihood of zero for a non-AGN whose S/N (V-band arm) is above 55. For galaxies with S/N values in the R-band and V-band arm less than the saturation values, we assign likelihoods as follows:

$$\begin{eqnarray}&&{L}_{T1}=\displaystyle \frac{{R}_{{sT}1}-{R}_{{fT}1}}{1-{R}_{{fT}1}-{R}_{{fT}2}}\end{eqnarray} \tag{ 5 }$$

$$\begin{eqnarray}&&{L}_{T2}=\displaystyle \frac{{R}_{{sT}2}-{R}_{{fT}2}}{1-{R}_{{fT}1}-{R}_{{fT}2}},\end{eqnarray} \tag{ 6 }$$

where L_sT1 and L_sT2 are the likelihoods for the galaxy to be a Type 1 and Type 2 AGN, respectively. We assign the saturation AGN fractions as their AGN likelihoods for galaxies that are not in our spectroscopic sample. The probability for a galaxy to be an AGN is the sum of the likelihoods for Type 1 and Type 2 AGNs. In this work, those galaxies whose spectra contaminated by telluric features around the AGN diagnostic emission lines are not in our spectroscopic sample. We assign them a flat likelihood of 0.3005, the saturation value for Type 1 and Type 2 AGNs. The galaxies identified as AGNs are assigned an AGN likelihood of 1.0. Examples of the likelihood assignments for the non-AGNs and galaxies with spectra contaminated by telluric absorption are given in Tables 3 and 4, respectively. These likelihoods can be used to statistically correct for both the inhomogeneity and incompleteness. These expressions for the likelihoods are only valid for the selection criteria used in this work, as specified in Sections 3.2 and 4. If the user chooses different criteria, e.g., different line S/N or BPT criteria, the homogeneity and completeness analyses should be repeated for the criteria used.

Table 3. Non-AGN Galaxies (Illustrative Extract)

TARGETNAME	R.A. (deg)	Decl. (deg)	z	Type	TMID	${\mathrm{SN}}_{{\rm{V}}}$	${\mathrm{SN}}_{{\rm{R}}}$	PROB1	PROB2	PROBAGN
g2359411_051614	359.92121	−5.27053	0.117294	⋯		6.0	10.2	0.0870	0.1811	0.2682
g2359433_540227	359.93050	−54.04086	0.084917	⋯		17.3	20.0	0.0646	0.1506	0.2152
g2359470_750411	359.94592	−75.06958	0.058782	Emi		7.9	7.0	0.0956	0.1740	0.2696
g2359485_455710	359.95192	−45.95289	0.065434	⋯		8.4	7.5	0.0944	0.1727	0.2672
g2359500_205008	359.95829	−20.83556	0.064346	⋯		9.2	9.0	0.0910	0.1710	0.2620
g2359553_123752	359.98029	−12.63114	0.104582	⋯		6.1	10.1	0.0872	0.1808	0.2680
g2359592_691411	359.99675	−69.23639	0.060147	⋯		5.2	7.0	0.0945	0.1820	0.2765
g2359390_484804	359.91246	−48.80111	0.065728	Emi		12.0	12.5	0.0831	0.1642	0.2473
g2359437_290936	359.93229	−29.15992	0.027535	Emi		9.2	9.6	0.0896	0.1712	0.2609
g2359547_301207	359.97804	−30.20192	0.030624	⋯		10.7	10.3	0.0882	0.1672	0.2554
g2359555_392832	359.98104	−39.47542	0.102470	⋯		10.0	15.9	0.0735	0.1719	0.2454
g0000086_062226	0.03601	−6.37399	0.021785	Emi	00000865-0622263	15.8	17.7	0.0705	0.1542	0.2247
g0000236_470108	0.09836	−47.01881	0.019984	⋯	00002363-4701076	23.0	10.3	0.0923	0.1277	0.2200
g0000586_333643	0.24409	−33.61195	0.023063	Emi	00005858-3336429	13.9	13.5	0.0810	0.1587	0.2397
g0001029_431950	0.26204	−43.33044	0.038783	⋯	00010289-4319496	21.6	24.0	0.0544	0.1378	0.1922
g0001175_530035	0.32276	−53.00967	0.032436	Emi	00011748-5300348	11.9	13.5	0.0804	0.1649	0.2453
g0001330_072610	0.38729	−7.43614	0.029474	Emi	00013295-0726099	24.8	25.1	0.0518	0.1271	0.1789

Note. TARGETNAME: 6dF ID; R.A.: R.A.; decl.: decl.; z: redshift; Type: galaxy type, "..." = non-emission-line galaxy, 'Emi'=galaxy with emission lines; TMID: 2MASS ID, indicate common galaxies from 2MRS; PROB1: likelihood to be a Type 1 AGN, i.e., L_sT1; PROB2: likelihood to be a Type 2 AGN satisfying the Kewley et al. (2001) criteria, i.e., L_sT2; and PROBAGN: likelihood to be an AGN, sum of L_sT1 and L_sT2. The full table also includes the emission-line widths, fluxes, and flux errors of all the non-AGN in the catalog.

Only a portion of this table is shown here to demonstrate its form and content. A machine-readable version of the full table is available.

Download table as:  DataTypeset image

We show the likelihood corrected AGN distribution, using the above Equations (3)–(6), across the sky and redshift in Figure 14. The AGN rates are now homogeneously distributed.

Zoom In Zoom Out Reset image size

We also explore the ratio of Type 1 (Sy1) to Type 2 (K01) AGN. The Type 1 to Type 2 AGN ratios vary widely across the sky and redshift as shown in Figure 15. This is expected since the Type 1 (Sy1) and Type 2 (K01) AGN fractions have different dependences on continuum S/N as seen in Figure 12. We correct the effect of the S/N in assigning AGN likelihood by using Equations (3)–(6) where the Type 1 and Type 2 AGN likelihoods are separately addressed. The statistical correction mostly eliminates the inhomogeneity in the Type 1/Type 2 ratio across the sky and in redshift, as shown in Figure 16.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

This catalog, like other optical AGN catalogs, is affected by dust obscuration. The only indicator of reddening available to us is the Balmer decrement, Hα/Hβ, which has a nominal value of 3 for unobscured AGNs and increases with obscuration. However, due to lack of absolute flux calibration, and Hα, Hβ located in two different arms, small values of Balmer decrements obtained from 6dF spectra are not reliable. Therefore, we cannot assess and correct for the inhomogeneity in this catalog due to dust obscuration.

We have constructed a southern-sky catalog of AGNs based on the optical spectroscopy of the 6dF galaxy survey. The catalog consists of 3109 broadline AGNs, and 12,156 narrowline AGNs that satisfy the Kauffmann et al. (2003) criteria, of which 3865 also satisfy the Kewley et al. (2001) criteria. We report emission-line widths, fluxes, flux errors, and S/Ns for all the galaxies in our spectroscopic sample, allowing users to customize the selection criteria. We perform an assessment of the completeness and homogeneity of our catalog across the sky and in redshift. Due to the inhomogeneity in the S/N, the AGN fractions are not evenly distributed. Especially at higher redshift, data with lower S/N are dominating the sample, where the detection rates rapidly drop down. In order to minimize the effect of continuum S/N on Type 1 (Type 2) AGN detection efficiency, the minimum requirement for the continuum S/N for a 6dF spectrum should be ≥55 in the R band (S/N ≥ 40 in the V band).

We provided means to correct for the S/N oriented inhomogeneity in our AGN catalog. An AGN likelihood is assigned to the objects with low S/N and those whose diagnostic lines are contaminated by the telluric absorptions. After these corrections, the AGN fractions are uniform across the southern sky and in redshift, making the catalog suitable for rigorous statistical analysis.

We thank the anonymous referee for the helpful comments. We are grateful to Heath Jones and Philip Lah for their help with the 6dF spectra. We thank Joseph Gelfand for his helpful comments on our manuscript. The research of G.R.F. was supported by National Science Foundation grant NSF-PHY-2013199. The research of I.Z. is supported by NYU Abu Dhabi Grant AD013. Y.C. acknowledges the support of NYU Abu Dhabi Grant AD013 and NYU Abu Dhabi Grant AD022.