Fast Outflowing Warm Absorbers in Narrow-line Seyfert 1 Galaxy PG 1001+054 Revealed by HST/COS Spectra

Active galactic nuclei (AGNs) are powered by the accretion of gas onto a supermassive black hole (SMBH), and AGN-driven feedback is generally considered important in the coevolution between AGN and its host galaxy over the cosmological timescale (e.g., Fabian 2012). The physical processes of feedback that could have an impact on the interstellar medium (ISM) and the intergalactic environment remain one of the most intensively studied subjects in the past two decades.

Outflows, either in the form of winds or jets, are important carriers of the energy output in feedback and are found to be prevalent in AGNs (e.g., King & Pounds 2015; Blandford et al. 2019; Veilleux et al. 2020; Silpa et al. 2022). They are common and frequently observed from  blueshifted absorption lines both in UV and X-ray bands (Bu & Yang 2021; Laha et al. 2021; Byun et al. 2022). It is currently believed that the possible origins of the outflow include accretion disk wind (Elvis 2000), torus wind (Krolik & Kriss 2001), board-line region (hereafter, BLR) clouds (Risaliti & Elvis 2010), and narrow-line region (hereafter, NLR) clouds (Kinkhabwala et al. 2002). The velocity of the outflow is usually several hundred to several thousand kilometers per second (McKernan et al. 2007); hence, it can propagate to a location of about 10 kpc (assuming a ∼10 million year lifetime of the AGN) away from the central black hole. In order to investigate the feedback efficiency of the AGN, it is necessary to quantify the physical properties of the outflows (Meena et al. 2021).

Sometimes the inferred velocity of the outflows detected via absorption of highly ionized metal lines can reach tens of thousands of kilometers per second. Such outflows are defined as ultrafast outflows (UFOs; Chartas et al. 2002; Reeves et al. 2003; Tombesi et al. 2010), which are likely launched at the scale of the innermost accretion disk. The mass outflow rates caused by these outflows can reach dozens of solar mass per year (Tombesi et al. 2013), implying large kinetic energies as expected in theoretical predictions of AGN feedback models (e.g., Hopkins & Elvis 2010; King 2010; King & Pounds 2015). Such energetic outflows may suppress the star formation in the host galaxy, prevent the gas around the galaxy from cooling (Krongold et al. 2007), and increase the abundance of intergalactic filaments (Hopkins & Elvis 2010).

Warm absorbers (WAs), characterized by narrow absorption lines and photoabsorption edges, also reveal the presence of the ionized phase outflow. They are detected in about 50% of Type I Seyfert galaxies either in the ultraviolet (UV) or in the X-ray band (Reynolds 1997; Crenshaw et al. 1999; Blustin et al. 2005). The WAs detected in the soft X-ray typically have higher ionization states compared to their UV counterparts (Lee et al. 2013; Laha et al. 2014; Fu et al. 2017). Compared to UFOs, WA outflows typically show much lower velocities, in the range of a few 100–1000 km s⁻¹, and a kinetic luminosity ≪1% of the bolometric luminosity of the AGNs (Blustin et al. 2005; Tombesi et al. 2013).

UV spectra typically have high spectral resolution allowing accurate dynamic measurement and cover line transitions of ions in low ionization states. In contrast, X-ray spectra cover more transitions from ions at various ionization states but are subject to lower resolution and signal-to-noise ratio (S/N; e.g., Kaspi et al. 2002). It is challenging to establish the one-to-one correspondence for the outflows between the UV and X-ray bands. The absorption column densities of outflows in these two bands can differ by hundreds of times (e.g., Ulrich 1988). As a result, outflows in these two bands may only partially overlap (Kriss 2004). Taking into account the rapid variations in luminosity, the approach adopted in recent years is to conduct simultaneous observations of certain sources (e.g., Mrk 509, Mrk 279, NGC 4051, 1H0419-577, NGC 3227; Costantini et al. 2007; Costantini 2010; Kriss et al. 2011; Di Gesu et al. 2013). A broad understanding of these WAs in the two bands has yet to emerge (Crenshaw et al. 2003).

The possible connection among the various types of ionized outflows was investigated in detail in Tombesi et al. (2013), where the comparison between the UFOs and the WAs suggests both types of ionized outflows belong to a single stratified outflow. Recent studies using high-resolution grating have detected UFOs in the soft X-rays (e.g., Ark 564, PDS 456, PG 1211+143, IRAS 17020+4544, Mrk 1044; Gupta et al. 2013; Longinotti et al. 2015; Nardini et al. 2015; Pounds et al. 2016b; Reeves et al. 2020; Krongold et al. 2021), typically with lower column densities and ionization parameters compared to those of the UFOs identified in the 6–9 keV range. Interestingly, the high-velocity UV counterparts to some of these UFOs have also been reported (Kraemer et al. 2012; Hamann et al. 2018; Kriss et al. 2018; Mehdipour et al. 2022) showing narrow and blueshifted absorption lines (e.g., HI Lyα, N v, C iv, etc.). Altogether, these strengthen the conjecture that they could be part of the same stratified multi-ionization outflow. Previous work further discovered a correlation between the wind outflow velocity and the hard X-ray luminosity of the AGN, suggesting that the UFOs could be consistent with a predominately radiatively driven wind (Matzeu et al. 2017; Mizumoto et al. 2021), arising in systems accreting at or close to the Eddington rate. Nevertheless, such a physical relationship and launching mechanism still remain unclear (see review by Longinotti 2020; Laha et al. 2021).

Over the past decade, Hubble Space Telescope (HST)/Cosmic Origins Spectrograph (COS) has performed highly sensitive high-resolution observations toward UV bright quasars, yielding abundant information of the intervening absorption systems, such as the gaseous halos and the circumgalactic medium of galaxies along the sight line to the quasar (e.g., Werk et al. 2013; Richter et al. 2016; Tumlinson et al. 2017; McCabe et al. 2021). Often spectral features associated with background quasars were not fully analyzed. This underexplored collection of archival spectra is a treasure trove for carrying out WA, UFO, and AGN outflow studies in general, which motivated our work.

A peculiar subgroup of AGNs characterized by high Eddington ratios (Boroson & Green 1992; Pounds & Vaughan 2000) are Narrow-line Seyfert 1 (NLS1) galaxies. Their other typical features include (1) the FWHM of hydrogen Balmer lines less than 2000 km s⁻¹ (Turner et al. 1999); (2) the line ratio of [O iii] λ5007 Å to Hβ less than 3 (Leighly 1999); (3) the high line ratio of Fe ii to Hβ (Mathur 2000); (4) generally steeper soft X-ray continuum slopes compared to other Seyfert 1 galaxies and rapid soft X-ray variability (Boller et al. 1996); (5) a strong infrared emission indicating active star formation (Moran et al. 1996). Indeed, several UFOs with high-velocity X-ray and UV absorbers are reported in NLS1 galaxies, e.g., NGC 4051 (Pounds & Vaughan 2012) and IRAS 17020+4544 (Mehdipour et al. 2022).

NLS1 galaxies with ionized outflows were selected based on the catalog of Rakshit et al. (2017), which provides a list of 11101 NLS1 galaxies identified from the Sloan Digital Sky Survey Data Release 12. This catalog is about 5 times larger than the number of previously known NLS1 galaxies. We retrieved the HST/COS observation of these sources and obtained a total of 72 NLS1 galaxies with UV observations. We identified one object with unpublished COS spectra that shows intriguing high-velocity absorption features, PG 1001+054, for a pilot study of the WA from the UV perspective.

PG 1001+054 (z = 0.16012, from NED ¹ ) is classified as an NLS1 galaxy, with an FWHM of Hβ ∼1740 km s⁻¹ (Wills et al. 2000). Meanwhile, the broad C iv and Lyα absorption lines in the UV spectrum also qualify it as a broad-absorption line (BAL) quasar (Brandt et al. 2000; Wang et al. 2000; Wills et al. 2000), a subclass with extremely weak X-ray emission. Wang et al. (2000) found that the observed UV line optical depth is much lower than expected from the X-ray absorbing column density in this source, based on a comparison between the ROSAT soft X-ray detection and the UV BALs. Despite the low S/N, the analysis of its XMM-Newton EPIC spectra (Schartel et al. 2005) found evidence for absorption through ionized material, modeled with a column density N_H of 19.2 × 10²² cm⁻² and an ionization parameter ξ of 542 erg s⁻¹ cm, respectively. Recent NuSTAR observations (Wang et al. 2022) investigated the nature of its X-ray weakness and suggested that X-ray obscuration by clumpy dust-free wind is sufficient to explain the variation of multiepoch X-ray data and the X-ray weakness.

In this work, we present the first detailed spectroscopic analysis and photoionization modeling of the high-resolution COS spectra of PG 1001+054. We identify high-velocity WAs (v ∼ 6700–8900 km s⁻¹), which demonstrates that high-resolution UV spectroscopy with HST can play a crucial role in feedback studies providing physical information, such as kinematics and ionization structure about the fast outflowing gas. Section 2 describes the observation and data reduction. In Section 3 we analyze the emission and absorption components in the spectra and prepare suitable models for them, using the same procedure in Zhang et al. (2015) and Fu et al. (2017). We present the fitting results in Section 4. In Section 5 we discuss and interpret the results of our modeling, and finally Section 6 gives a brief summary of our conclusions.

PG 1001+054 was observed by HST/COS in 2014 June (PI: J. Bregman) with an exposure time of 7 ks. It was observed as a target in a large sample to study missing baryons in nearby dwarf galaxies and has not been studied in detail individually. The log of the observations is listed in Table 1. The COS observations consist of eight segments utilizing two gratings, G130M and G160M, which are centered at 1300 Å and 1600 Å, respectively. This yields spectra fully covering a wavelength range of 1135 Å to 1795 Å (Osterman et al. 2011), with a resulting medium spectral resolution R ≡ λ/Δλ from 16,000 to 21,000.

The archival HST/COS data of PG 1001+054 were retrieved from the Mikulski Archive for Space Telescopes, ² fully calibrated and processed with the latest COS calibration pipeline CAL COS (version 3.3.11). The wavelengths of expected local ISM lines and geocoronal lines are used to verify that the calibrated wavelength is accurate. We use IDL routines described in Danforth et al. (2010) to process flat-fielding, alignment, and coaddition. Eight observations are merged with exposure weighting, and the final spectrum's S/N covers a range of 15–25 per resolution element (0.07 Å, or 17 km s⁻¹). Using the IDL toolkit Package for Interactive Analysis of Line Emission ³ (Kashyap & Drake 2000), the COS flux spectrum is finally converted into the format commonly used in X-ray studies. Pulse-height amplitude files and the corresponding response (RSP) files that convolved with G130M and G160M line-spread functions are generated following the above process. Figure 1 presents the final COS spectrum of PG 1001+054.

Zoom In Zoom Out Reset image size

The processed spectrum is analyzed and fitted by the Interactive Spectral Interpretation System ⁴ (version 1.6.2; Houck 2002), which is a programmable, interactive tool to explore the physics of the COS spectrum.

Taking the cosmological redshift of PG 1001+054 in account, we first identify the typical UV emission and absorption lines in the spectrum based on AtomDB ⁵ (version 3.0.9; Foster et al. 2012). Narrow absorption lines caused by the diffuse ISM lines have also been identified and masked to exclude interference when fitting the spectrum. Next we utilize the photoionization code XSTAR (version 2.54a; Kallman 2001) to model the emitting and absorbing plasma photoionized by the radiation from the central accretion disk. To generate XSTAR model grids, several parameters need to be set, for example, the luminosity of the AGN, the hydrogen nucleus density, the metal abundances (usually set to solar values), the spectral energy distribution (SED) file, the temperature, and the turbulent velocity. We set the column density N_H and the ionization parameter ξ = L_ion/(nr²) as intrinsically free parameters, where L_ion is the luminosity in the 1–1000 Ryd energy range, n is the hydrogen nucleus density, and r is the distance to the ionizing source. The parameters for XSTAR table models are listed in Table 2. We use the b parameter in XSTAR to produce the corresponding FWHMs in the observed line profiles, where $b\,=\mathrm{FWHM}/2\sqrt{(\mathrm{ln}2)}\approx \mathrm{FWHM}/1.665$.

Considering the UV radiation could be substantially reduced by dust extinction, we obtain extinction-corrected radiation using the extinction curve formula proposed by Gordon et al. (2009), which covers the wavelength range from 910 Å to 3.3 μm. Parameters R_v = A_V /E(B − V) = 3.1 (Cardelli et al. 1989) and A_V = 0.042 (available from NED) are adopted in the case of PG 1001+054. To construct the SED file of PG 1001+054, we collect the available data from NED supplemented by the data from our dust-extinction-corrected UV spectrum. PG 1001+054 is also known as a bright low redshift quasar, with a bolometric luminosity L_bol = 1.35 × 10⁴⁵ erg s⁻¹ (Pennell et al. 2017). The ionizing luminosity of PG 1001+054 is obtained from the SED file as L_ion = 2.14 × 10⁴⁴ erg s⁻¹.

The intrinsic UV radiation of PG 1001+054 generally comes from the accretion disk, BLR, and NLR. A power-law model is used to fit the multiple blackbody emission in the accretion disk empirically. As for the emission from BLR and NLR clouds, we need to generate XSTAR models to fit these components. We take the plausible assumption that the clouds in the BLR and NLR have approximately virial speed, i.e., $v\propto \sqrt{{\rm{G}}M/R}$ (M is the black hole mass, and R represents the distance to the center) following the method in our previous work (Zhang et al. 2015; Fu et al. 2017). In this case, the width of the emission line in the spectrum is closely linked to the radial velocity dispersion of different groups of photoionized clouds associated with the BLR and NLR. Each BLR or NLR component is assumed to have similar physical conditions, such as density, radial distance, and temperature and thus can be described by one photoionization component. We caution that recent work on spatially resolved NLR outflows in local AGNs clearly reveals variations in ionization and density (Revalski et al. 2021), and we do not further interpret on the model fits of emission line.

These emission lines are first adequately modeled to facilitate identification of absorption systems. Following Zhang et al. (2015) and Fu et al. (2017), we model the emission lines using XSTAR. Some previous work adopted other approaches for modeling the emission line, such as multiple Gaussians (e.g., He et al. 2017; Meena et al. 2021), piecewise function, or spline fit (e.g., Veilleux et al. 2022). To generate those photoionization components in XSTAR, we fit the strong emission lines with Gaussian models to measure the FWHM of each line and derive the Doppler broadening parameter or turbulent velocity. In the COS spectrum of PG 1001+054, the best candidate lines for profile decomposition are Lyα, N v doublet, and Si iv doublet, given their high S/N relative to other lines. The five lines are fitted jointly according to their rest-frame wavelengths. The five lines share the same radial velocity and FWHM, and the flux ratio of these doublets is fixed at 2:1 for an optically thin case. Two groups of Gaussians with FWHM values of ∼4031 and 1069 km s⁻¹ are needed. These quite large FWHM values indicate that the two emission components are likely associated with the BLRs although FWHM ∼ 1000 km s⁻¹ could be possible for a highly turbulent NLR. We tentatively associate the two components as BLR clouds (hereafter, BLR1 and BLR2). Two XSTAR table models are generated for them. The density, column density, and temperature of the BLR model are set to 10¹⁰ cm⁻³, 10²³ cm⁻², and 15,000 K, respectively, following the value in Zhang et al. (2015). The metallicities are set to the solar values, and we leave the ionization parameter and the redshift as free parameters. The measured redshift by XSTAR will allow us to identify offsets from the systemic velocity due to outflow or inflow. After the initial fitting, residual emission represented by one Gaussian component is added to the model. The reduced chi-square of the global best fit is 1.64.

We estimate the outflow velocities of WAs through the Lyα, N v, and Si iv doublet absorption troughs in the COS spectrum. By fitting these absorption lines with Gaussians, we identify three kinematic components (hereafter, WA I, WA II, and WA III) with blueshifted velocities of −8931, −6761, and −8201 km s⁻¹. The FWHMs of these three WAs are about 1047, 850, and 479 km s⁻¹, respectively. Since thermal broadening and turbulent broadening can make a significant impact on the absorption lines, we can calculate the v_turb of each WA according to $\mathrm{FWHM}=2\sqrt{\mathrm{ln}2}\sqrt{{v}_{\mathrm{th}}^{2}+2{v}_{\mathrm{turb}}^{2}}$ (Zhang et al. 2015). v_th is the thermal velocity, which is about ${v}_{\mathrm{th}}=13\sqrt{T/10,000A}$ km s⁻¹, where T is the temperature and A represents the atomic number. Typically the temperatures of low ionization WAs are about 10⁵ K, and the mean value of the carbon and nitrogen atomic numbers is adopted as the A value. All of the outflow velocities and turbulent velocity v_turb values are listed in Table 3. The separate model components we fit in PG 1001+054 spectra are shown in Figure 2.

Zoom In Zoom Out Reset image size

Table 3. Intrinsic Lines Fitted with Gaussian in the COS Spectrum

Ion(λrest)	fa	flux( × 10−4) b	λobs	Velocity c	FWHM
Å		Photons−1 cm−2	Å	km s−1	km s−1
Emission Lines from the BLR
Lyα (1215.67)	0.41617	84.67	1414.03	⋯	⋯
		177.33	1410.25	⋯	⋯
N v (1238.82)	0.15553	45.05	${1441.96}_{-0.25}^{+0.47}$	${914.50}_{-58.58}^{+112.12}$	${4031.33}_{-115.59}^{+86.81}$
		10.05	1437.11 ± 0.02	−17.91 ± 5.21	${1069.14}_{-9.31}^{+21.42}$
N v (1242.81)	0.07781	22.53	1445.59	⋯	⋯
		5.03	1441.73	⋯	⋯
Si iv (1393.76)	0.52771	62.46	1621.17	⋯	⋯
		2.59	1616.84	⋯	⋯
Si iv (1402.77)	0.26285	31.23	1631.66	⋯	⋯
		1.30	1627.30	⋯	⋯
Absorption Lines from Warm Absorbers
Lyα (1215.67)	0.41617	14.97	1382.89	⋯	⋯
		726.00	1377.06	⋯	⋯
		2.04	1374.11	⋯	⋯
N v (1238.82)	0.15553	88.19	1409.24	⋯	⋯
		21.49	1403.29	⋯	⋯
		7.58	${1400.28}_{-0.06}^{+0.03}$	$-{8931.12}_{-14.52}^{+7.26}$	${479.01}_{-34.44}^{+41.46}$
N v (1242.81)	0.07781	44.10	${1413.77}_{-0.02}^{+0.04}$	$-{6761.60}_{-4.82}^{+9.65}$	${1046.68}_{-17.17}^{+17.99}$
		10.74	${1407.80}_{-0.08}^{+0.03}$	$-{8201.69}_{-19.30}^{+7.24}$	${850.14}_{-32.71}^{+34.95}$
		3.79	1404.78	⋯	⋯
Si iv (1393.76)	0.52771	4.26	1585.49	⋯	⋯
		0.10	1578.79	⋯	⋯
		0.10	1575.40	⋯	⋯
Si iv (1402.77)	0.26285	2.13	1595.74	⋯	⋯
		0.05	1589.00	⋯	⋯
		0.05	1585.59	⋯	⋯

Notes.

^a f is the oscillator strength. ^b The strong and weak line flux ratio of the doublets is in an optically thin case of 2:1. ^c Five lines in one component have the same radial velocity and FWHM because they are fitted jointly according to their rest wavelength relations.

Download table as:  ASCIITypeset image

We adopt the typical value of the density and temperature in these WAs as 10⁴ cm⁻³ and 10⁵ K (Fu et al. 2017). The metallicities are set to the solar values. The column density N_H, the ionization parameter ξ, and the redshift z of the absorbing gas are left as free parameters. Our assumption here is that each WA component is only sampling one distinct slab in the outflowing material along our line of sight. Finally, the XSTAR models of these WAs are generated, and the parameters used for these XSTAR models are listed in Table 2.

We also attempt to derive constraints of the covering factor by the WAs. The compact accretion disk is generally considered fully covered by the WAs, whereas the BLR may not be completely covered by the WAs. Therefore, when fitting the COS spectrum, for accretion disks, the covering factor is 1. For BLRs, it is determined by spectral fitting, and the XSPEC model partcov is used to mimic this effect.

Figure 1 shows the observed COS spectrum and the best-fit model including continuum, line emission, and absorption. We manually added a Gaussian absorption model component to account for the local broad Lyα absorption and a Gaussian emission component for the local O i doublet (1302 Å and 1306 Å) emission. The observed spectrum with complex line profiles appears well fitted by these physical components. The photon index of the underlying power law is Γ = 3.20. A total of three components of WAs are used to describe the identified absorption lines in the observed COS spectrum. Table 4 lists the parameters of the best-fit WAs in PG 1001+054. Synthetic spectral models illustrating these components are shown in Figure 3. All three WAs have noticeably high blueshifted velocities, ranging from ∼6600 to ∼8900 km s⁻¹. WA I has the lowest velocity, column density, and ionization parameter among the three WAs, suggesting that it may be the farthest away from its central black hole. WA II and III have similar higher ionization states and larger column densities, and their velocities are higher than WA I. In Figure 4 we show the detailed contribution of the absorption lines in each WA (Lyα, N v, and Si iv doublet) to the total multicomponent absorption features. As shown in Figure 5, the best-fit XSTAR models for Lyα, N v, and Si iv absorption in PG 1001+054 are compared to the observed COS spectrum.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Table 4. Parameters of the Emission Line Clouds and Identified Warm Absorbers

	logξ	NH	Redshift	Covering
	(erg s−1 cm)	(×1019 cm−2)		Factor
Emission
BLR I	${3.00}_{-0.01}^{+0.09}$	⋯	${0.160765}_{-0.000051}^{+0.000007}$	⋯
BLR II	0.90 ± 0.01	⋯	${0.159983}_{-0.000001}^{+0.000014}$	⋯
Intrinsic Absorption
WA I	0.65 ± 0.02	${6.48}_{-0.18}^{+0.14}$	${0.137902}_{-0.000018}^{+0.000007}$	0.91
WA II	1.90 ± 0.02	${79.31}_{-4.17}^{+4.01}$	${0.132926}_{-0.000029}^{+0.000025}$	0.12
WA III	2.05 ± 0.02	${32.98}_{-3.42}^{+1.57}$	${0.130239}_{-0.000029}^{+0.000026}$	0.02

Download table as:  ASCIITypeset image

5.1. Comparison with Previous Work

Being an X-ray weak BAL quasar (Brandt et al. 2000; Luo et al. 2014; Wang et al. 2022), there are few previous studies on the absorbers of PG 1001+054. Schartel et al. (2005) found that the column density N_H and ionization parameter ξ of outflows in this source are 19.2 × 10²² cm⁻² and 542 erg s⁻¹ cm, respectively, from its XMM-Newton observations. However, the column density N_H and ionization parameters ξ of the absorber that we measured in the UV band are much lower than the X-ray derived values. It is plausible that the absorbers detected in the two bands are at different locations.

For absorption systems, it is notoriously difficult to obtain the location of absorbing gas along the line of sight. We attempt to estimate the distance of the WAs away from the central black hole according to ξ = L_ion/(nr²). L_ion is the luminosity in the 1–1000 Ryd energy range, n is the hydrogen number density, and r is the distance to the ionizing source. Without absorption lines from excited states to constrain the density, we cannot determine the exact value of n_H for these WAs in this source; nevertheless, the n_H values for similar absorbers in other AGNs are generally larger than 10³ cm⁻³ (Gabel et al. 2005; Arav et al. 2013, 2015; Miller et al. 2018; Aalto et al. 2020; Arav et al. 2020). Due to the higher velocity of the outflowing absorbers in this source than those of typical WAs, it is likely that the n_H may also be slightly larger; hence a lower limit of n_H as 10³ cm⁻³ is adopted. On the other hand, previous work studying BAL/mini-BAL outflows frequently take r = 1 pc as a placeholder radial distance to derive relevant properties (Hamann et al. 2019 and references therein). Following this approach, we find the corresponding n_H is between 2 × 10⁵ and 5 × 10⁶ cm⁻³ for our outflows. Using these values, we obtained that a plausible range for the location of these absorbers is between 1 and 73 pc. Table 5 lists the detailed information of the radial distance of these WAs from the nucleus. The location of these WAs implies that these WAs may be possibly linked to the ISM beyond the torus scale.

Table 5. The Derived Properties of the Identified WAs in PG 1001+054

	logξ	NH	vout	ne	r	${\dot{M}}_{\mathrm{out}}$	$\dot{{E}_{k}}$	$\dot{{E}_{k}}/{L}_{\mathrm{Edd}}$
	(erg s−1 cm)	(×1019 cm−2)	(km s−1)	(cm−3)	(pc)	(M⊙ yr−1)	(erg s−1)
WA I	0.65	6.48	−6660.73	103 − 5 × 106	1–73	0.05–4.00	7.36 × 1041 − 5.57 × 1043	0.013% − 0.967%
WA II	1.90	79.31	−8152.49	103 − 3 × 105	1–17	0.11–1.87	2.26 × 1042 − 3.92 × 1043	0.039% − 0.681%
WA III	2.05	32.98	−8958.02	103 − 2 × 105	1–15	0.01–0.12	2.08 × 1041 − 3.03 × 1042	0.004% − 0.053%

Download table as:  ASCIITypeset image

It is also worth noting that the location of the absorbers observed in the UV band appears to be slightly farther from the central black hole than those observed in the X-ray band. Wang et al. (2000) also reached the same suggestion, finding that the X-ray absorbers require column densities of at least a few times 10²² cm⁻² in this source, which are much larger than those inferred from the UV absorption lines. They compared the ROSAT X-ray data and the UV absorption and emission lines and suggested that the observed UV line optical depth is much lower than expected from the X-ray absorbing column density. Our results are consistent with their conclusions.

In a broader context, the outflowing UV absorber identified in PG 1001+054 can be compared to a well-studied sample of UV/X-ray absorbers in local AGN and PG quasars. For example, Laha et al. (2014) carried out a homogeneous analysis of WAs in X-rays (WAX) in 26 Seyfert galaxies using the XMM-Newton spectra and performed linear regression fits for the WA parameters that could be compared as typical values. The seminal work by Tombesi et al. (2013) also provided a sample of soft X-ray WAs of Type 1 Seyfert galaxies. In addition, from the literature we compiled, the derived properties for a list of NLS1 sources similar to PG 1001+054 that have reported UFO measurements, either in the X-ray or in the UV. These include IRAS 17020+4544 (Mehdipour et al. 2022; Sanfrutos et al. 2018), Mrk 1044 (Krongold et al. 2021), PG 1211+143 (Pounds et al. 2016a), 1H1934-063 (Xu et al. 2022), PG 1448+273 (Laurenti et al. 2021), 1H 0707-495 (Kosec et al. 2018), IRAS 13224-3809 (Jiang et al. 2022), and Mrk 590 (Gupta et al. 2015). In Figure 6 we show the distribution of different outflow parameters for these sources together with our measurements for PG 1001+054. Overall, the data points are rather scattered in the parameter space, with the exception of the ionization parameter ($\mathrm{log}\xi$) versus column density ($\mathrm{log}{N}_{{\rm{H}}}$). Note that no fit for correlation analysis is intended here. Laha et al. (2014) previously identified there is such a correlation. Figure 6(a) shows that the UV absorber in PG 1001+054 is typical of the WAs in ionization and column density, but in terms of outflow velocity, it is between UFOs (v ∼ 0.1c) and other typical WAs (v ∼ 1000 km s⁻¹). The significantly higher outflow velocity (v ∼ 7000–9000 km s⁻¹) implies that a fast outflow is clearly present in this NLS1. This is consistent with the expectation that the launching mechanism of the UFOs is related to the radiatively driven wind (Matzeu et al. 2017; Mehdipour et al. 2022) and that NLS1 galaxies have overall high Eddington ratios.

Zoom In Zoom Out Reset image size

Together with the detection of high ionization absorbers in the X-ray for PG 1001+054 (Schartel et al. 2005), we suggest that the fast outflowing UV absorber seen in the COS spectrum is probably part of a multiphase outflow. In the X-ray obscuration scenario, Wang et al. (2022) suggest a powerful high-density wind launched from the accretion disk. Most likely the low resolution and low S/N X-ray spectra due to its nature of an X-ray weak quasar prohibited firm detection of any X-ray UFO in this source. Figure 6 illustrates that the parameters of outflows span a wide range in ionization states, column densities, and velocities, which are consistent with the characteristics of stratified winds. Several AGNs were detected with multicomponent UFOs in terms of both ionization and velocity (e.g., IRAS 17020+4544, Mehdipour et al. 2022; PDS 456, Reeves et al. 2018), showing Lyα UFO features besides the X-ray absorption signatures (e.g., Fe XXV, Fe XXVI, O VII, etc). As discussed in Krongold et al. (2021), such a structure is likely produced by the ambient ISM shocked by outflow launched at the accretion disk scale (King & Pounds 2015).

5.2. Mass and Energy Outflow Rate

In this work, we analyze the HST/COS spectra of the NLS1 galaxy PG 1001+054 and report on the discovery of fast outflowing UV absorbers. We fit the high-resolution UV spectrum of PG 1001+054 and perform photoionization models, taking into account the physical components of local gas absorption, local dust extinction, Comptonized corona emission of the AGN, blackbody emission from the accretion disk, BLR emission, and absorption due to intrinsic WAs. The main findings are summarized as follows.

• 1.  
  The UV outflow is seen as narrow and blueshifted absorption lines of Lyα, N v, and Si iv, with velocities in the range of ∼7000–9000 km s⁻¹. The presence of three WAs can explain well the significant absorption lines in the COS spectrum, and we derive their physical properties. A fast outflow is clearly present in this NLS1 and consistent with the expectation that launching mechanism of outflows is related to radiatively driven wind and that NLS1 galaxies have overall high Eddington ratios.
• 2.  
  The possible location of WAs can be estimated to range between ∼1 pc and ∼73 pc away from the central black hole, which implies that these WAs may originate at or beyond the torus scale. The UV absorber in PG 1001+054 is typical of WAs in ionization and column density but shows significantly higher outflow velocity than other WAs (v ∼ 1000 km s⁻¹). Together with previous detection of high ionization absorbers in the X-ray for PG 1001+054 (Schartel et al. 2005), we suggest that the fast outflowing UV absorber seen in the COS spectrum is probably part of a multiphase outflow. The parameters of outflows are consistent with the characteristics of stratified winds. Such a structure is likely produced by the ambient ISM shocked by outflow launched at the accretion disk scale.
• 3.  
  We estimate a total mass outflow rate of the three WAs as ${\dot{M}}_{\mathrm{out}}\,\approx \,$ 0.2–6 M_⊙ yr⁻¹. The ratio of total $\dot{{E}_{k}}$ of the WA outflows over L_Edd is ≈0.1%–1.7%, which could potentially sufficiently influence the evolution of the host galaxy when compared to the values in some theoretical models for efficient AGN feedback.

We sincerely thank the anonymous referee for the critical reading and helpful suggestions. We thank Dr. Shuinai Zhang and Mouyuan Sun for beneficial discussion and Li Xue for technical assistance in computing. We also thank Jianfeng Wu, Bin Luo and Defu Bu for helpful advice. We acknowledge support by the NSFC grants (U1831205, 12033004, 12221003, 12103041), and the science research grants from CMS-CSST-2021-A06 and CMS-CSST-2021-B02. This research is based on observations made with the NASA/ESA Hubble Space Telescope obtained from the Space Telescope Science Institute, which is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS 526555. These observations are associated with HST program 13347.

Facility: HST (COS). -