The Awakening Beast in the Seyfert 1 Galaxy KUG 1141$+$371 I.

We use the standard pipeline NUPIPELINE v.0.4.6 in HEASOFT v.6.26.1 to reduce the NuSTAR data. The calibration file version is v.20181030. A circular region with radii of 50 arcsec is used to extract source spectra. The background spectra are extracted from the remaining regions on the same chip. The spectra are generated using the NUPRODUCTS tool. The FPMA and FPMB spectra are both grouped to have a minimum signal-to-noise of 6 and to oversample the energy resolution by a factor of 3. The FPM spectra are modelled over 3–60 keV.

A long-term X-ray lightcurve of KUG 1141 is shown in the top panel of Fig. 2. The observed flux is calculated in the overlapping 0.5–7 keV band of Swift XRT and XMM-Newton EPIC-pn. As shown in the figure, the X-ray flux from KUG 1141 increases by almost one order of magnitude from $\log(F_{\rm X}/$erg cm^-2 s^-1)=-11.79 in 2007 to $\log F_{\rm X}=-10.92$ in 2017 (hereafter the X-ray flux values are reported in units of erg cm^-2 s^-1). The latest observation in 2019 (obs 12) shows that KUG 1141 still remains in a high flux state with $\log F_{\rm X}=-11.08$ since 2017.

The X-ray softness of KUG 1141, which is defined as the flux ratio between the 0.5–2 keV and 2–7 keV bands, is shown in the second panel of Fig. 2. The variability of the softness ratio indicates that the X-ray spectrum of KUG 1141 shows a ‘softer-when-brighter’ pattern, and is the softest in 2017 when the X-ray flux reaches the highest level. Obs 12 in 2019 shows a slightly harder continuum compared to obs 11.

Long-term UV and optical lightcurves \textcolorblackof the AGN in KUG 1141 are shown in the third and fourth panels of Fig. 2. Note that the XMM-Newton OM was operated with only UVW1 and UVM2 filters during obs 5. The rest of the flux values are given by the Swift UVOT observations in the archive. The following conclusions can be drawn:

blackFirst, the UV and optical emission from the AGN of KUG 1141 shows simultaneous flux increase in the period of 2009–2019. For instance, UVW2 band flux has increased by more than one order of magnitude from 0.12 mJy in 2009 to 1.23 mJy in 2019; B band flux has increased by a factor of 3 from 0.47 mJy in 2009 to 1.53 mJy in 2019.

Second, we find that the UV flux increases by a larger factor than the optical flux. In the \textcolorblackfifth panel of Fig. 2, we show of the UVW2 and V flux ratio as an example. Note that the V and UVW2 bands are respectively the longest and the shortest UVOT wavelengths. The flux ratio between these two bands indicates the steepness of the UV–optical continuum emission. As shown in Fig. 2, the UV–optical flux ratio is the highest in 2009 when the source is in a low optical/UV flux state; the ratio is the lowest in 2017 and 2019 when the source is in a high optical/UV flux state.

Third, we follow the same approach as in Vignali et al. (2003) to calculate $\alpha_{\rm OX}$, which is defined as $0.384\log(f_{\rm 2keV}/f_{2500})$. $f_{\rm 2keV}$ and $f_{2500}$ are respectively the flux density in units of erg cm^-2 s^-1 Hz^-1 at 2 keV and 2500 Å in the rest frame. 2500 Å in the rest frame of KUG 1141 corresponds to 2598 Å in the observer’s frame, which is within the wavelength range of the UVW1 filter (2600 Å) on UVOT. By approximating the UVW1 flux as the 2500 Å flux, we calculate $\alpha_{\rm OX}\approx\alpha^{\prime}_{\rm OX}=0.384\log({f_{\rm 2keV}/f_{\rm UVW1}})$. The values of $\alpha^{\prime}_{\rm OX}$ are shown by the right y-axis of the sixth panel in Fig. 2. $\alpha^{\prime}_{\rm OX}$ is relatively constant before 2015, and the source becomes X-ray-weak since 2017. The relative weakness of X-ray emission in KUG 1141 might be related to enhanced reflection or absorption (e.g. Gallo, 2006).

We first model the three EPIC spectra of obs 5 with an absorbed power-law model cutoffpl (Model 1). Due to the lack of simultaneous hard X-ray observation, we fix the high energy cutoff parameter $E_{\rm cut}$ at 500 keV. Model 1 is able to fit the EPIC spectra very well with $\chi^{2}_{\rm red}=0.95$. The best-fit parameters are shown in Table 2, and the corresponding ratio plot is shown in Fig. 4. \textcolorblackWe did not find obvious evidence for narrow emission feature or absorption edge features in the iron band.

In order to test for possible soft excess emission, we add an additional bbody model (Model 2), which improves the fit by only $\Delta\chi^{2}=6$ with 2 more parameters. See the left panel of Fig. 4 for comparison of the fits. Only an upper limit of the temperature parameter is obtained ($kT$<0.2 keV). The normalization parameter of the bbody model is not constrained. A contour plot of $\chi^{2}$ distribution on the $kT$ vs. normalization parameter plane is shown in the right panel of Fig. 4. The normalization of the bbody component is consistent with a very low value, e.g. <$1\times 10^{-6}$, within a 3$\sigma$ uncertainty range. When $kT$ is at a very low value, e.g. <0.07 keV, the fit is no longer sensitive to $kT$ due to the lower limit of the energy coverage of EPIC.

We also test for any extra absorption along the line of the sight towards KUG 1141 by adding an additional tbnew model at the source redshift. We only obtain an upper limit of the column density (a 90% confidence range of $N_{\rm H}<3\times 10^{20}$ cm^-2).

To sum up, we conclude that the X-ray spectra of KUG 1141 during obs 5 are consistent with a Galactic-absorbed power law. There is no/little evidence for soft excess emission or additional line-of-sight absorption.

We first model the disc thermal emission of KUG 1141 using the diskbb model. This model assumes a thin disc temperature profile of $T\propto r^{-0.75}$ and has two parameters: the inner temperature of the disc ($kT_{\rm in}$) and the normalization parameter. The normalization parameter is defined as $(R^{\prime}_{\rm in}/D_{\rm 10})^{2}\cos i$, where $R^{\prime}_{\rm in}$ is the ‘apparent’ inner radius of the disc in km, $D_{\rm 10}$ is the source distance in 10 kpc, and $i$ is the inclination angle of the disc. See Section 5.2 for more discussion concerning the normalization parameter of diskbb. Note that both $kT_{\rm in}$ and the normalization parameter can change the total flux of the model.

Secondly, we calculate the Comptonisation spectrum of the hot coronal region using the convolution model simpl (Steiner et al., 2009). The simpl model calculates a power law-shaped Comptonisation spectrum that self-consistently accounts for the fraction of scattered radiation. The spectrum of the seed photons is the diskbb model as described above. In this way, we are able to constrain the strength of the disc thermal emission by consistently considering the scattering process, and thus estimate the inner radius of the disc ($R_{\rm in}$). Such a method was previously used to measure $R_{\rm in}$ in BH transients when the sources are not necessarily in a thermal-dominant state (e.g. Steiner et al., 2009). The free parameters are the scattering fraction $f_{\rm scatt}$ and the photon index of the power-law continuum ($\Gamma$).

The total model is tbnew * zdust * zmshift * ( simpl * diskbb ) in the XSPEC format. Such a model can describe the SEDs of most epochs very well except obs 12. An additional weak bbody model is used to model the soft excess shown in obs 12 as demonstrated in Section 4.2.1. The best-fit SED models are shown in Fig. 8, and the best-fit parameters are shown in Table 10 and Fig. 9. In the top panel of Fig. 9, we also show the absorption-corrected broad band flux calculated between 0.01 eV–100 keV using our best-fit AGN component. The following conclusions can be drawn from our SED modelling:

1) An increase of the disc temperature together with an increase of the broad band flux is clearly seen according to our SED modelling. \textcolorblackFor instance, the best-fit inner temperature of the disc is $kT_{\rm in}=2.0\pm 0.2$ eV during obs 4; a higher temperature of $kT_{\rm in}=4.8^{+0.5}_{-0.4}$ eV is required for obs 11. See the bottom middle panel of Fig. 8 for comparison of these two epochs. The increase of the disc temperature is able to explain the variability of the optical–UV continuum: the UV emission is more sensitive to the increase of the disc temperature than the optical emission, and thus increases by a larger factor than the optical emission as shown in fifth panel of Fig. 2.

2) \textcolorblackDespite an increase of the disc temperature, the normalization of the diskbb model decreases from $\approx 2.6\times 10^{10}$ in 2009 to $\approx 5\times 10^{9}$ in 2019. Note that, again, an increase of either $kT_{\rm in}$ or the normalization parameter can increase the flux of the model. The decrease of the normalization parameter suggests a decreasing inner radius of the disc as this parameter is proportional to $R^{2}_{\rm in}$. See Section 5.2 for more discussion concerning the inner radius of the disc.

3) The scattering fraction parameter ($f_{\rm scatt}$) shows the number fraction of the disc seed photons that are up-scattered to the X-ray band in the coronal region and gives a physical interpretation of the X-ray–UV ratio shown in Fig. 2. $f_{\rm scatt}$ is the highest during obs 11, where the X-ray flux and the X-ray–UV ratio are both shown to be the highest among all the epochs. It is interesting to note that the best-fit $f_{\rm scatt}$ for obs 12 is similar to the values for obs 4-10 when the source is in a much lower UV and optical flux state.

A change in the coronal region might be a possible explanation: during the high flux state during obs 11, a different corona, e.g. of a larger size in which more disc seed photons are up-scattered, might exist. Therefore, a high fraction of accretion power is released in the form of non-thermal X-ray emission rather than the thermal emission from the disc during obs 11. The corona during obs 12 might however be similar to those during obs 4-10 despite a higher mass accretion rate. Our studies suggest that the X-ray–UV ratio/$f_{\rm scatt}$ do not necessarily show strict correlation with mass accretion rate in an individual AGN.

4) The photon index of the X-ray continuum ($\Gamma$) changes with $f_{\rm scatt}$. This suggests that the optical depth of the corona is sensitive to not only the mass accretion rate as shown by statistical studies of AGN surveys (e.g. Brightman et al., 2013) but also the energy interplay between the disc and the corona.

Assuming $L_{\rm Bol}\approx L_{\rm 0.01eV-100keV}$ and a BH mass of $10^{8}M_{\odot}$⁶⁶6 Oh et al. (2015) estimated $\log(M_{\rm BH})=7.99\pm 0.06$ (statistical error) for KUG 1141 based on the study of H$\alpha$ emission line by following the method in Greene & Ho (2005). Systematic errors of this method can lead to a larger uncertainity range, such as the relation between $\rm FWHM_{\rm H\alpha}$ and $\rm FWHM_{\rm H\beta}$ (Greene & Ho, 2005) and the uncertain geometry of the broad-line region (Kaspi et al., 2000; McLure & Dunlop, 2004). , we estimate the Eddington ratio of KUG 1141 in the past decade. For example, \textcolorblacka broad band flux of $F_{\rm 0.01eV-100keV}=2\times 10^{-11}$ erg cm^-2 s^-1 during obs 4 corresponds to an Eddington ratio of $\lambda_{\rm Edd}=0.6\%$. Our best-fit SED model suggests that the bolometric luminosity of KUG 1141 increases from $\approx 0.6\%$ of the Eddington limit in 2009 to $\approx 3.2\%$ in 2019. A similar fraction of mass accretion rate increase has also been seen in other flaring AGN (e.g. NGC 1566, Parker et al., 2019).

From the normalization parameter of the diskbb model, we are also able to estimate the inner radius of the disc. This parameter is defined as $(R^{\prime}_{\rm in}/D_{\rm 10})^{2}\cos i$ as introduced above, where $R^{\prime}_{\rm in}$ is the ‘apparent’ inner radius. We assume a correction factor of 0.412 (Kubota et al., 1998) to convert $R^{\prime}_{\rm in}$ to the real inner radius $R_{\rm in}$. For instance, \textcolorblackassuming a viewing angle of $i=10^{\circ}$, the best-fit normalization parameter of diskbb for obs 4 is $2.6\times 10^{10}$ corresponding to an inner radius of $R_{\rm in}\approx 8$ $r_{\rm g}$. Based on our calculations, the inner disc radius of KUG 1141 decreases from 8 $r_{\rm g}$ in 2009 to 3 $r_{\rm g}$ in 2017 and 2019 assuming $i=10^{\circ}$. An assumption for a higher viewing angle leads to a larger estimated value: $R_{\rm in}$ decreases from 20 $r_{\rm g}$ in 2009 to 8 $r_{\rm g}$ in 2019 assuming $i=80^{\circ}$. However, KUG 1141 is a Sy1 galaxy which is unlikely to be viewed from an edge-on angle according to the standard model of AGN (e.g. Antonucci, 1993) assuming the torus plane and the optical-emitting disc plane share the same inclination.

blackIn conclusion, we find evidence for a possible decreasing inner radius simultaneously with an increasing mass accretion rate in KUG 1141, and the value $R_{\rm in}$ has decreased by a factor of approximately 2.7. The assumptions on viewing angle and black hole mass do not affect the change of parameters we observe but only the absolute values.

Runco et al. (2016) studied the SDSS and Keck observations of KUG 1141 in the optical band, which were taken respectively in 2005 and 2009. These optical spectra of KUG 1141 are consistent with a typical Sy1. Meanwhile, these two observations are mostly consistent, which suggests that the disc of KUG 1141 does not show large variability in the period of 2005-2009.

By analysing the observations after 2009, we find that KUG 1141 has shown a steady increase of optical and UV flux. The luminosity of KUG 1141 increases from 0.6% of the Eddington limit in 2009 to 3.2% in 2017. The latest UVOT observation suggests that UV and optical flux of KUG 1141 is still increasing.

Detailed SED modelling shows that the inner radius of the disc ($R_{\rm in}$) in KUG 1141 decreases by a factor of approximately $2.7$. \textcolorblackThe exact value of $R_{\rm in}$ depends on the assumption for the disc inclination angle: $R_{\rm in}$ decreases, for example, from 8 $r_{\rm g}$ (20 $r_{\rm g}$) in 2009 to 3 $r_{\rm g}$ (8 $r_{\rm g}$) in 2019 assuming $\theta=10^{\circ}$ ($80^{\circ}$). Saxton et al. (2015) estimated the filling time of a truncated thin disc in the framework of the standard thin disc model (Shakura & Sunyaev, 1973). They found that, for example, it takes more than 800 years to refill a truncated disc with $R_{\rm in}=20$ $r_{\rm g}$ around a $10^{8}$ $M_{\odot}$ BH as in KUG 1141 based on viscous timescales. Therefore, the change of $R_{\rm in}$ in KUG 1141 is much faster than expected.

The mechanisms behind a quick boost of mass accretion rate as in KUG 1141 are also still unclear. It might be related to the instability of a gas pressure-dominated disc at a low Eddington ratio (Saxton et al., 2015), the propagation process in the disc (e.g. Ross et al., 2018) or tidal disruption events (TDEs, Rees, 1988).

In the case of TDEs, it is interesting to note that theoretically the rise time for a TDE to reach the peak luminosity is less than 1 year, assuming a 1 $M_{\odot}$ star disrupted by a $10^{8}$ $M_{\odot}$ BH as in KUG 1141$+$371 and the tidal radius is twice the periastron radius (De Colle et al., 2012). Indeed similar conclusions have been found in TDEs that have been very well monitored before their peak luminosity (e.g. Bade et al., 1996; Cenko et al., 2012; Arcavi et al., 2014). However, KUG 1141 has shown a steady flux increase in the past decade.

Moreover, there are three typical types of TDE X-ray spectra: 1) strong thermal emission (e.g. ASASSN-14li, Grupe et al., 2015); 2) very soft power-law continuum (e.g. $\Gamma>2$ in XMMSL2 J144605.0$+$685735, Saxton et al., 2019); 3) hard power-law continuum (e.g. $\Gamma\approx 1.6$ in Swift J164449.3$+$573451, Burrows et al., 2011). The spectra of KUG 1141 do not agree with the former two types but only the third type of TDEs, which is only found in radio-loud galaxies with significant jet emission in X-rays. This is however not the case for KUG 1141 where there is no obvious evidence for radio emission from KUG 1141 in the VLA FIRST Survey (Wadadekar, 2004). Therefore, the increase of the mass accretion rate in KUG 1141 is unlikely due to TDEs, or KUG 1141 is at least not consistent with typical TDEs we have observed.

In this section, we discuss possible connection between KUG 1141 and BH transients in outburst. BH transients are binary systems where the central accretors are stellar-mass BHs. The existence of two different flux states in these sources have been realised for decades (e.g. Oda et al., 1971). Their X-ray spectra can change from a soft spectrum characterised by a strong thermal component to a hard spectrum characterised by a power-law component. Such a transition is commonly seen during an outburst of a transient and shows a ‘Q’-shaped pattern in the X-ray hardness-intensity diagram (e.g. see a MAXI HID of GX 339$-$4 in Jiang et al., 2019a).

Similarly, we show the HID for KUG 1141 in Fig. 10. The circles in the figure represent Swift observations, and the square represents the XMM-Newton observation in 2009. The ‘intensity’ is the total absorption-corrected AGN flux in the 0.01 eV–100 keV band given by the best-fit SED models. The thermal emission from the disc in AGN is in the optical and UV bands. Therefore, we show X-ray–UV ratio, which we calculate in Section 3, in the diagram as ‘hardness’ instead of X-ray hardness as for BH transients.

As shown in Fig. 10, KUG 1141 starts from the bottom right corner of the diagram in 2009 when the UV, optical and X-ray emission is all at a low flux level. Then the total luminosity of KUG 1141 starts increasing in 2014 while the X-ray–UV ratio still remains consistent. When the source reaches the highest X-ray flux state in 2017, a lower X-ray–UV ratio is found. The latest observation in 2019 shows that the UV and optical flux is still increasing. The current state of KUG 1141 is near the top left corner of the HID. Such a pattern is very similar to the ‘Q’-shaped HID of many BH transients in outburst.

blackThe latest Swift and NuSTAR observation of KUG 1141 (obs 12) suggests a hard X-ray continuum of $\Gamma=1.7$ at a high accretion rate. The combination of a hard X-ray power-law emission and strong disc thermal emission makes obs 12 of KUG 1141 a similar case to the intermediate state seen in some outbursts of BH transients. During these states, a stellar-mass BH transient often shows a modest mass accretion rate of $\lambda_{\rm Edd}\approx 1\%$, and the disc thermal component and the non-thermal power-law component of $\Gamma<2$ make similar contribution to the X-ray emission (e.g. GRS 1716$-$249, Jiang et al., 2020a).

blackBesides, the outbursts of BH transients can last for various timescales, e.g. weeks (e.g. MAXI 1659$+$152, Negoro et al., 2010) or years (e.g. Swift J1753.5$-$0127, Soleri et al., 2013). The corresponding lengths for a flaring AGN scaled by BH mass are much longer than observable timescales, e.g. $\approx 10^{6}$ years. Therefore, sources like KUG 1141 can show an outburst with a significant increase of accretion rate similar to a BH transient on very short timescales are very intriguing.

Last but not least, a persistent radio jet is often seen during the hard state observations of BH transients (e.g. Fender et al., 2005) while a quenching of the radio emission is observed during the transition to the soft state (e.g. Fender et al., 1999). On contrary, KUG 1141 was not observed by the VLA FIRST survey (Wadadekar, 2004), which was taken before the transition shown in Fig. 10 started. If KUG 1141 indeed shows a BH transient-like outburst, our hypothesis would predict simultaneous variability in the radio emission. A radio monitoring program for KUG 1141 in future will be able to answer the question.

blackChanging-look AGN are rare cases of AGN, where the optical continuum flux increases or decreases and the broad emission lines appear or disappear within short time-scales. In previous studies, Noda & Done (2018) suggested that some ‘changing-look’ AGN may have strong radiation or magnetic pressure in the disc, which may shorten the state transitions in AGN that are similar to BH transients. Ruan et al. (2019) showed that the combined $\alpha_{\rm OX}$–$\lambda_{\rm Edd}$ evolution of two ‘changing-look’ AGN show a similar shape as KUG 1141 does in Fig. 10 with a similar increase of $\lambda_{\rm Edd}$.

blackAll of these properties make KUG 1141 a similar case to changing-look AGN. However, it is also important to note that KUG 1141 does not seem to change ‘look’, e.g. Sy2 to Sy1, along with the boost of accretion rate: the XMM-Newton observation taken in 2009 when the source was in a low-$\lambda_{\rm Edd}$ state was able to rule out a Compton-thick scenario that is commonly seen in Sy2 AGN (Risaliti et al., 1999); during the high-$\lambda_{\rm Edd}$ state, e.g. obs 12 in 2019, the X-ray spectra of KUG 1141 still remained unobscured. Besides, the Keck observation in the optical band presented in Runco et al. (2016) showed that KUG 1141 already had a typical Sy1 spectrum in 2009 during the low-$\lambda_{\rm Edd}$ state⁷⁷7Our follow-up optical observations taken by Lijiang Observatory during the high-$\lambda_{\rm Edd}$ state will be presented in the second paper of this series..

We find strong evidence for soft excess below 1 keV only in obs 12 when KUG 1141 is in the high flux state ($\lambda_{\rm Edd}\approx 3\%$). By modelling the soft excess emission with a phenomenological blackbody model (bbody), we obtain $kT=0.1$ keV, which is similar to the value of a typical Sy1 (e.g. Gierliński et al., 1999).

Noda & Done (2018) suggests that a rising soft excess should be seen when an AGN goes through a ‘changing-look’ phase with a modest mass accretion rate, e.g. $\lambda_{\rm Edd}\approx 1\%$. However, we are only able to obtain an upper limit of the soft excess emission in obs 11 taken in 2017 when the X-ray flux is the highest and the continuum is the softest. This is contrary to the example of Mrk 1018 given by Noda & Done (2018).

There are two most popular explanations for the origin of the soft excess emission in Sy galaxies: warm corona and relativistic disc reflection.