An X-ray Luminosity-dependent “Changing-look” Phenomenon in UGC 3223

J. Wang Guangxi Key Laboratory for Relativistic Astrophysics, School of Physical Science and Technology, Guangxi University, Nanning 530004, Peopleʼs Republic of China Key Laboratory of Space Astronomy and Technology, National Astronomical Observatories, Chinese Academy of Sciences, Beijing 100101, Peopleʼs Republic of China D. W. Xu Key Laboratory of Space Astronomy and Technology, National Astronomical Observatories, Chinese Academy of Sciences, Beijing 100101, Peopleʼs Republic of China School of Astronomy and Space Science, University of Chinese Academy of Sciences, Beijing, Peopleʼs Republic of China J. Y. Wei Key Laboratory of Space Astronomy and Technology, National Astronomical Observatories, Chinese Academy of Sciences, Beijing 100101, Peopleʼs Republic of China School of Astronomy and Space Science, University of Chinese Academy of Sciences, Beijing, Peopleʼs Republic of China

Abstract

The nature of the rare “Changing-look” (CL) phenomenon in active galactic nuclei (AGNs) is still under debate at current stage. We here present Swift/XRT and UVOT follow-up observations of UGC 3223, a newly discovered repeat CL-AGN with type transitions of $\mathrm{S1.5\rightarrow S2\rightarrow S1.8}$ occurring in a period of about 30 years. By comparing the values previously reported in the ROSAT All-sky Survey and in the second Swift-XRT Point Source catalog, we clearly find that the X-ray flux tightly follows the optical spectral transition, in which a spectral type closer to a Seyfert 1 type is associated with a higher X-ray flux. An invariable X-ray spectral shape is, however, found in the CL phenomenon of the object. An extremely low Eddington ratio of $\sim 2\times 10^{-4}$ can be obtained from the X-ray luminosity for its Seyfert 2 state, which suggests a favor of the disk-wind broad-line region model in explaining the CL phenomenon. A variation of the total UV emission is not revealed when compared to the previous GALEX NUV observation, since the UVOT images indicate that $\sim 90$ % UV emission comes from the intensive star formation in the host galaxy.

galaxies: Seyfert — galaxies: nuclei — X-rays: galaxies — ultraviolet: galaxies

^†^†facilities: Swift(XRT and UVOT)^†^†software: HEASOFT, XSPEC (Arnaud 1996)

1 Introduction

A present hot topic is the nature of the rare “changing-look” active galactic nuclei (CL-AGNs) that manifest themselves a spectral type transition between type I, intermediate type, and type II within a timescale of an order of years to decades. Up to now, there are only $\sim 100$ CL-AGNs identified by multi-epoch spectroscopy (e.g., MacLeod et al. 2010, 2016; Shapovalova et al. 2010; Shappee et al. 2014; LaMassa et al. 2015; McElroy et al. 2016; Ruan et al. 2016; Runnoe et al. 2016; Parker et al. 2016; Gezari et al. 2017; Sheng et al. 2017, 2020; Stern et al. 2018; Yang et al. 2018; Wang et al. 2018, 2019, 2020; Frederick et al. 2019; Trakhtenbrot et al. 2019; Yan et al. 2019; Ai et al. 2020; Graham et al. 2020; Kollatschny et al. 2018, 2020). Although almost all the CL-AGNs are identified by a dramatic variation of their Balmer emission lines, Guo et al. (2019) recently reported the first Mg II CL-AGN: SDSS J152533.60+292012.1. The discovery of the CL-AGNs challenges the widely accepted AGN paradigm, not only in the orientation-based unified model (e.g., Antonucci 1993), but also in the standard disk model in terms of the viscosity crisis (e.g., Lawrence 2018 and references therein).

The physical origin of CL-AGNs is still under debate, although some progress has been achieved in past a couple of years. Based on the light echo in middle infrared (Sheng et al. 2017) and spectropolarimetry (e.g., Hutsemekers et al. 2019), there is evidence suggesting that a variation of the accretion power of a supermassive black hole (SMBH) is a plausible explanation of the rare CL phenomenon. The statistical study in MacLeod et al. (2019) and case study in Wang et al. (2020) recently argue that the CL phenomenon can be possibly understood in the context of the disc-wind broad-line region (BLR) model (e.g., Elitzur & Ho 2009; Nicastro 2000; Elitzur & Shlosman 2006), in which a critical Eddington ratio ( $L/L_{\mathrm{Edd}}$ , $L_{\mathrm{Edd}}=1.26\times 10^{38}M_{\mathrm{BH}}/M_{\odot}\ \mathrm{erg\ s^{-1}}$ is the Eddington luminosity) of $\sim 10^{-3}$ ( $10^{-6}$ ) is predicted for an (dis)appearance of broad emission lines.

Ruan et al. (2019) recently found an anti-correlation between UV-to-X-ray spectral index and $L/L_{\mathrm{Edd}}$ in some CL-AGNs, which suggests a linkage between the accretion state transition occurring in X-ray binaries (XRBs) and the CL phenomenon observed in AGNs. This linkage was subsequently reinforced by Ai et al. (2020 and references therein) who identified a relation between spectral type transition and variation of X-ray hardness in a small sample of CL-AGNs, i.e., a well-established V-shape correlation between X-ray hardness and $L/L_{\mathrm{Edd}}$ .

Wang et al. (2020) recently identified a new nearby CL-AGN, UGC 3223 (1RXS J045909.6+045831, $z=0.015621$ ), with a repeat type transition. So far, repeat type transitions have been only identified in a few nearby Seyfert galaxies and quasars (e.g., Parker et al. 2019; Marin et al. 2019 and references therein). Based on the spectrum taken by Stirpe (1990) in 1987, our spectroscopic observations of the object over the course of 18 yr enable us to witness its type transition from 1.5 $\rightarrow$ 2.0 $\rightarrow$ 1.8 over 32 yr. Our latest spectroscopy in 2020/02 suggests the object is now still in the “turn-on” state with a Seyfert 1.8 spectrum.

In this paper, we report Swift/XRT and UVOT follow-up observations, taken at the “turn-on” state, of this interesting object, and explore the origin of its CL phenomenon by examining the variation of its X-ray and UV emission by comparing with previous results extracted from literature. The paper is organized as follows. Section 2 describes Swift/XRT and UVOT follow-up observations and data reductions. The variation of its X-ray and UV emission is studied in Section 3. Section 4 shows the conclusion and discussion. A $\Lambda$ CDM cosmology with parameters $H_{0}=70\mathrm{km\ s^{-1}\ Mpc^{-1}}$ , $\Omega_{m}=0.3$ and $\Omega_{\Lambda}=0.7$ is adopted throughout the paper.

2 Observations and Data Reductions

We proposed X-ray and ultraviolet follow-up observations for the object in 2020 by using the Neil Gehrels Swift Observatory (Gehrels et al. 2004) X-ray telescope (XRT) and Ultraviolet/Optical Telescope (UVOT) . The object was targeted three times (ObsID=00049202002, 00049202003 and 00049202004) by XRT and UVOT simultaneously on 2020 March 31, April 6 and April 10. The log of these observations is shown in Table 1.

2.1 XRT X-ray Spectrum

We reduced the XRT data taken in the Photon Counting (PC) mode by the HEASOFT version 6.27.2, along with the corresponding CALDB version 20190910. All the three XRT exposures were at first stacked to enhance the signal-to-noise (S/N) ratio. The source spectrum was then extracted from the stacked image by a circular region with a radius of 23.5″. The corresponding background spectrum was extracted from an adjacent region free of any sources. The corresponding ancillary response file was produced by the task xrtmkarf. In fact, the XRT count rates in 0.3-10keV are $(4.77\pm 0.75)\times 10^{-2}$ , $(4.51\pm 1.37)\times 10^{-2}$ and $(5.14\pm 0.80)\times 10^{-2}\ \mathrm{count\ s^{-1}}$ for the three exposures, which suggests a stable count rate level within the period.

The stacked X-ray spectrum was then modeled by XSPEC (v12.11, Arnaud 1996) in terms of four models. The best fitting and best-fit parameters are shown in Figure 1 and Table 2, respectively. All the uncertainties quoted in the table correspond to a 90% significance level.

We first modeled the spectrum by a simple model expressed as $wabs*powerlaw$ over the 0.3-10keV in terms of the C-statistic (Cash 1979; Humphrey et al. 2009; Kaastra 2017), where the Galactic hydrogen column density is fixed to be $N_{\mathrm{H}}=8.1\times 10^{20}\ \mathrm{cm^{-2}}$ (Kalberla et al. 2005). Although the fitting is acceptable with a C-statistic of 1.283, it returns a quite hard spectrum with a photon index of $\Gamma=0.82^{+0.29}_{-0.30}$ . This unusually hard spectrum implies either an existence of strong excess absorption, or else a significant starburst contribution, although the latter is unlikely (see Section 4).

In order to include a potential local absorption, a model of $wabs*zwabs*powerlaw$ was then used to reproduce the observed spectrum. However, the best-fit returns a zero value of the intrinsic hydrogen column density $N_{\mathrm{H}}$ when it is set as a free parameter in the modeling. A fixed value of $N_{\mathrm{H}}$ was therefore adopted in our modeling. With the fixed intrinsic $N_{\mathrm{H}}$ , a softer powerlaw with $\Gamma=1.24^{+0.33}_{-0.34}$ can be obtained from the best fit. We determine the $N_{\mathrm{H}}$ from either H I 21cm observation or extinction determined from the Balmer decrement as follows.

We estimate the intrinsic $N_{\mathrm{H}}$ from the total H I mass $M_{\mathrm{HI}}$ as $N_{\mathrm{H}}=M_{\mathrm{HI}}/\pi ab$ , where $a$ and $b$ are the major and minor radius of the host galaxy, respectively. Given the standard way of estimating $M_{\mathrm{HI}}$ from the observed H I 21cm line flux (Roberts 1962), the intrinsic $N_{\mathrm{H}}$ can be calculated from

N_{\mathrm{H}}=\frac{1.218\times 10^{24}}{a\arcsec b\arcsec}\frac{F^{\prime}_{\mathrm{HI}}}{\mathrm{Jy\ km\ s^{-1}}}\ \mathrm{cm^{-2}}

(1)

where both $a\arcsec$ and $b\arcsec$ are in units of arcsecond, and $F^{\prime}_{\mathrm{HI}}$ is related with the observed 21cm line flux as $F^{\prime}_{\mathrm{HI}}=(a/b)^{0.12}F_{\mathrm{HI}}$ (Giovanelli et al. 1994). Using the measured $F_{\mathrm{HI}}=3.05\ \mathrm{Jy\ km\ s^{-1}}$ (Courtois et al. 2009) enables us to estimate intrinsic $N_{\mathrm{H}}$ to be $\sim 1.3\times 10^{21}\ \mathrm{cm^{-2}}$ . We alternatively estimate $N_{\mathrm{H}}$ from the standard relationship between $N_{\mathrm{H}}$ and extinction $A_{V}$ : $N_{\mathrm{H}}/A_{V}\sim(1.8-2.2)\times 10^{21}\mathrm{atoms\ cm^{-2}\ mag^{-1}}$ , where $A_{V}$ can be determined from the the observed Balmer decrement of the narrow emission lines. With the standard Case B recombination, an extinction of $A_{V}=1.08$ mag can be inferred from the spectroscopy reported in Wang et al. (2020), which finally leads to a $N_{\mathrm{H}}=1.9-2.4\times 10^{21}\ \mathrm{cm^{-2}}$ .

The spectrum can be also reproduced by a model of $wabs*zpcfabs*powerlaw$ by including a neutral partially covering absorber. This modeling results in a little better C-statistic of 0.868 and a photon index of $\Gamma=2.06^{+0.69}_{-1.09}$ that is close to the mean value of AGNs of $\sim 1.8$ .

Taking into account the intensive star formation activity occurring in the host disk (see Section 2.2), the spectrum is found to be best reproduced by a model including an emission from hot, diffuse gas (Raymond & Smith 1977), i.e., $wabs*zwabs*(powerlaw+raymond)$ . In this case, we obtain a photon index of $\Gamma=2.13\pm 0.66$ and a plasma electronic temperature of $kT_{\mathrm{e}}=0.05^{+0.02}_{-0.01}$ keV, along with a C-statistic of 1.023.

2.2 UVOT Images

The UVOT images were reduced by the HEASOFT and corresponding CALDB version 20170922. Again, we at first co-add the multi-exposures to enhance the S/N ratio for each filter. The corresponding exposure maps of the co-added images were generated by the uvotexpmap task. Figure 2 displays the co-added UVOT images in the three used filters, i.e., $uvm2$ , $uvw1$ and $uvw2$ from left to right. The object is spatially resolved in all the UVOT filters. One can see from the three images that the UV emission is dominantly contributed from a ring and a spot in the host disk, rather than the nucleus, which indicates an intensive star formation activity occurring in the host disk. We argue that the revealed ring structure in UV with an extent of about 30″ is most likely due to the large scale dust lane that was identified in the HST WFPC2 image by Deo et al. (2006).

An elliptical aperture with a radius of 36.4″, which is twice the effective radius reported in Bai et al. (2015), was adopted to measure its total brightness. Again, the corresponding background level was determined from an adjacent region free of any sources. The Galactic extinction is corrected by using the Galactic Reddening Map (Schlegel et al. 1998) and the Galactic extinction curve in Cardelli et al. (1989). The results of our photometry in the AB magnitude system are $uvm2=16.31\pm 0.04$ mag, $uvw1=16.10\pm 0.04$ mag and $uvw2=16.58\pm 0.03$ mag, where the reported magnitude errors include both statistical and system uncertainties.

Refer to caption — Table 1: Log of observations carried out by Swift/XRT and UVOT.

ObsID	Obs start time	XRT exposure in PC mode	UVOT exposure	UVOT filters
		seconds	seconds
(1)	(2)	(3)	(4)	(5)
00049202002	2020-03-31 21:19:34	857.5	845.1	$uvm2,uvw1,uvw2$
00049202003	2020-04-06 23:56:35	268.3	261.7	$uvm2,uvw2$
00049202004	2020-04-10 06:08:35	818.3	808.0	$uvm2,uvw1,uvw2$

Parameter	Value	Units	Description
(1)	(2)	(3)	(4)
Model 1 - $wabs*powerlaw$
$\Gamma$	$0.82^{+0.29}_{-0.30}$		Powerlaw index
$F(\mathrm{2-10keV})$	$3.44^{+1.38}_{-1.00}\times 10^{-12}$	$\mathrm{erg\ s^{-1}\ cm^{-2}}$	Unabsored flux
Cash statistics	7.70/7=1.100
Model 2 - $wabszwabspowerlaw$
$N_{\mathrm{H}}$	0.24	$10^{22}\mathrm{cm^{-2}}$	Local column density (fixed)
$\Gamma$	$1.24^{+0.33}_{-0.34}$		Powerlaw index
$F(\mathrm{2-10keV})$	$2.86^{+1.18}_{-0.85}\times 10^{-12}$	$\mathrm{erg\ s^{-1}\ cm^{-2}}$	Unabsored flux
Cash statistics	10.77/7=1.539
Model 3 - $wabszpcfabspowerlaw$
$\eta_{\mathrm{H}}$	$2.74_{-1.95}^{+2.03}$	$10^{22}\mathrm{cm^{-2}}$	Local equivalent column density
$f$	$0.89_{-0.61}^{+0.09}$		Dimensionless covering fraction
$\Gamma$	$2.06^{+0.69}_{-1.09}$		Powerlaw index
$F(\mathrm{2-10keV})$	$3.13^{+1.33}_{-0.93}\times 10^{-12}$	$\mathrm{erg\ s^{-1}\ cm^{-2}}$	Unabsored flux
Cash statistics	4.23/5=0.868
Model 4 - $wabszwabs(powerlaw+raymond)$
$N_{\mathrm{H}}$	$2.00^{+0.99}_{-0.58}$	$10^{22}\mathrm{cm^{-2}}$	Local column density (free)
$\Gamma$	$2.13\pm 0.66$		Powerlaw index
$kT_{\mathrm{e}}$	$0.05^{+0.02}_{-0.01}$	keV	Temperature of hot, diffuse gas
$F(\mathrm{2-10keV})$	$2.88^{+0.61}_{-0.53}\times 10^{-12}$	$\mathrm{erg\ s^{-1}\ cm^{-2}}$	Unabsored flux of the powerlaw
Cash statistics	4.09/4=1.023

Mission	Date of Obs.	$F_{\mathrm{0.1-2.4keV}}$	$F_{\mathrm{0.3-10keV}}$	$HR1_{\mathrm{RASS}}$	$HR2_{\mathrm{RASS}}$	$HR1_{\mathrm{Swift}}$	$HR2_{\mathrm{Swift}}$	Sp. type	Reference
		$10^{-12}\mathrm{erg\ s^{-1}}$
(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)
ROSAT	1990-1991	15.50		$0.93\pm 0.03$	$0.28\pm 0.07$			S1.5	1
Swift	2013-2014		$0.34\pm 0.08$			$0.295^{+0.400}_{-0.300}$	$0.350^{+0.262}_{-0.204}$	S2	2
Swift	2020.04	$0.86^{+0.16}_{-0.14}$	$3.86^{+1.36}_{-1.00}$	$0.953\pm 0.233$	$0.664\pm 0.204$	$0.357\pm 0.170$	$0.351\pm 0.117$	S1.8	this work

An X-ray Luminosity-dependent “Changing-look” Phenomenon in UGC 3223

Abstract

1 Introduction

2 Observations and Data Reductions

2.1 XRT X-ray Spectrum

2.2 UVOT Images

3 Analysis and Results

3.1 X-ray Emission

3.2 UV Emission

4 Conclusion and Discussion

4.1 Star Formation Contribution

4.2 Physics of CL Phenomenon in UGC 3223

References