Swift monitoring of NGC 4151: Evidence for a Second X-ray/UV Reprocessing

TL;DR

This study uses Swift observations of NGC 4151 to analyze X-ray and UV/optical variability, revealing complex reprocessing behavior inconsistent with simple models and supporting a two-stage reprocessing scenario.

Contribution

It provides detailed, multi-band light curves and evidence for a two-stage reprocessing model in NGC 4151, challenging the traditional lamp-post paradigm.

Findings

Hard X-ray bands are strongly correlated with no lag.

UV/optical bands lag X-rays by 3-4 days.

UV leads optical by 0.5-1 day.

Abstract

Swift monitoring of NGC 4151 with ~6 hr sampling over a total of 69 days in early 2016 is used to construct light curves covering five bands in the X-rays (0.3-50 keV) and six in the ultraviolet (UV)/optical (1900-5500 A). The three hardest X-ray bands (>2.5 keV) are all strongly correlated with no measurable interband lag while the two softer bands show lower variability and weaker correlations. The UV/optical bands are significantly correlated with the X-rays, lagging ~3-4 days behind the hard X-rays. The variability within the UV/optical bands is also strongly correlated, with the UV appearing to lead the optical by ~0.5-1 day. This combination of >~3 day lags between the X-rays and UV and <~1 day lags within the UV/optical appears to rule out the "lamp-post" reprocessing model in which a hot, X-ray emitting corona directly illuminates the accretion disk, which then reprocesses the energy in the UV/optical. Instead, these results appear consistent with the Gardner & Done picture in which two separate reprocessings occur: first, emission from the corona illuminates an extreme-UV-emitting toroidal component that shields the disk from the corona; this then heats the extreme-UV component which illuminates the disk and drives its variability.

Tables

Table 1. Table 1 Monitoring Information

(1)	(2)	(3)	(4)	(5)	(6)
	Central	Wavelength/	Number	Sampling	$F_{\mathrm{var}}$
Band	$\lambda$ (Å)	Energy Range	of Points	Rate (day)	(%)
BAT	0.45	15–50 keV	69	1.00	18.6
X4	1.8	5–10 keV	319	0.22	34.6
X3	3.5	2.5–5 keV	319	0.22	41.6
X2	7	1.25–2.5 keV	319	0.22	17.4
X1	20	0.3–1.25 keV	319	0.22	9.1
uvw2	1928	1650–2250 Å	254	0.28	6.1
uvm2	2246	2000–2500 Å	252	0.23	5.7
uvw1	2600	2250–2950 Å	273	0.26	5.4
u	3465	3050–3900 Å	276	0.22	6.0
b	4392	3900–4900 Å	319	0.22	3.9
v	5468	5050–5800 Å	310	0.23	2.4

Table 2. Table 2 UVOT Dropout Testing

(1)	(2)	(3)	(4)	(5)	(6)
Filter	Num	Dropouts	Num	Masked	Final
	Obs		Masked	Non-drop	Data
uvw2	308	57	54	2	254
uvm2	455	62	80	19	252aa
uvw1	322	38	49	11	273
u	320	14	44	31	276
b	319	2	0	0	319
v	310	4	0	0	310

Table 3. Table 3 UVOT Data

(1)	(2)	(3)	(4)
MJD	Flux	Error	Filter
57438.0447	5.767	0.080	uvw2
57438.3637	5.805	0.080	uvw2
57438.4962	5.755	0.080	uvw2
57438.6966	5.878	0.081	uvw2
57439.0400	5.788	0.080	uvw2

Table 4. Table 4 XRT Data

(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
MJD	X1 Flux	X1 Error	X2 Flux	X2 Error	X3 Flux	X3 Error	X4 Flux	X4 Error
57438.0435	0.132	0.024	0.073	0.018	0.328	0.038	0.430	0.044
57438.3631	0.108	0.019	0.100	0.018	0.236	0.028	0.322	0.033
57438.4963	0.134	0.020	0.090	0.016	0.304	0.030	0.358	0.033
57438.6955	0.192	0.032	0.082	0.021	0.436	0.048	0.405	0.046
57438.8985	0.150	0.021	0.071	0.014	0.361	0.032	0.408	0.034

Table 5. Table 5 BAT Data

(1)	(2)	(3)
MJD	BAT Flux	BAT Error
57438.5	0.00510	0.00124
57439.5	0.00591	0.00094
57440.5	0.00380	0.00157
57441.5	0.00555	0.00099
57442.5	0.00703	0.00080

Table 6. Table 6 uvw2 CCF Results

(1)	(2)	(3)	(4)	(5)	(6)
Band	$r_{\max }$	${\tau }_{\mathrm{med}}$ (days)	${\tau }_l$ (days)	${\tau }_u$ (days)	Sig. (%)
BAT	0.46	-7.42	-13.49	-3.08	77.2
X4	0.64	-3.58	-4.04	-3.22	88.6
X3	0.68	-3.39	-3.72	-3.11	90.8
X2	0.56	-2.28	-3.10	-1.58	83.1
X1	0.33	-3.74	-8.26	-1.55	68.8
uvw2	1.00	0.00	-0.25	0.24	$>$99.9
uvm2	0.97	0.01	-0.21	0.26	$>$99.9
uvw1	0.95	0.02	-0.24	0.29	$>$99.9
u	0.95	0.61	0.33	0.88	$>$99.9
b	0.89	0.83	0.49	1.15	$>$99.9
v	0.82	0.96	0.50	1.43	$>$99.9

Table 7. Table 7 X3 CCF Results

(1)	(2)	(3)	(4)	(5)	(6)
Band	$r_{\max }$	${\tau }_{\mathrm{med}}$ (days)	${\tau }_l$ (days)	${\tau }_u$ (days)	Sig. (%)
BAT	0.75	-0.42	-1.65	4.19	99.9
X4	0.92	-0.10	-0.24	0.04	$>$99.9
X3	1.00	0.00	-0.14	0.14	$>$99.9
X2	0.57	1.35	0.89	1.78	99.2
X1	0.34	1.07	-8.31	3.35	93.8
uvw2	0.68	3.40	3.12	3.72	91.4
uvm2	0.67	3.64	3.30	4.07	89.8
uvw1	0.64	3.68	3.23	4.09	87.4
u	0.60	3.63	3.22	4.17	85.1
b	0.58	3.13	2.67	3.69	88.5
v	0.56	4.16	3.28	5.25	91.0

Equations

$$r={\left(X\frac{k{\lambda }_0}{hc}\right)}^{4/3}{\left[\left(\frac{GM}{8\pi \sigma }\right)\left(\frac{L_{\mathrm{Edd}}}{\eta c^2}\right)(3+\kappa ){\dot{m}}_{\mathrm{E}}\right]}^{1/3}$$

$$r=ct=0.09{\left(X\frac{\lambda }{1928\mathring{\mathrm{A}}}\right)}^{4/3}{M}_8^{2/3}{\left(\frac{{\dot{m}}_{\mathrm{Edd}}}{0.10}\right)}^{1/3}\text{lt-dy}$$

Full text

1. Introduction

Although the quantity and quality of observational data on active galactic nuclei (AGN) has vastly improved over the past few decades, the standard model of the physical structure of the central engine has remained largely unchallenged. The fundamental picture of urrounding a supermassive black hole (SMBH) was first proposed by Lynden-Bell (1969). The model of an optically thick, geometrically thin accretion disk was first proposed by Shakura & Sunyaev (1973) in the context of stellar-mass black holes. Galeev et al. (1979) added magnetic reconnection in a corona above the disk in order to explain the observed hard X-ray emission from AGN. This predicts that the corona can directly illuminate and heat the outer disk (e.g., Frank et al. 2002), leading to the so-called “lamp-post” or “reprocessing” model. Note that in this paper the use of the term “lamp-post” does not require that the X-ray source must be a point source; instead, we only require that it is small relative to the ultraviolet and optical (UV/optical) emitting disk, extending above and below the disk so that it directly illuminates the disk.

A clear prediction of this model is that flux variations in the X-ray emitting corona will be seen in the UV/optical emission from the disk. Measurement of the interband X-ray/UV temporal lag and smoothing can then be used to estimate the size and structure of the disk. This technique, known as reverberation mapping (RM; Blandford & McKee 1982; Peterson 1993; Peterson 2014), has been used for decades in a different context to constrain the size and physical characteristics of the broad emission-line region (BLR; Peterson 1997). This model predicts a clear relation between lag ($\tau$) and wavelength ($\lambda$) as the variations from the smaller, hotter inner disk are expected to precede those from the larger, cooler outer disk regions, scaling as $\tau\propto\lambda^{4/3}$ (e.g., Cackett et al. 2007).

Application of RM to the corona/disk system has been more difficult than to the BLR because the sizes (and thus the lags) are much smaller ($\lesssim$1 day). Nonetheless, this “accretion disk RM” approach has been repeatedly attempted because of the potential large reward: information on the size and structure of the central engines of AGN that cannot be probed by any other method except gravitational lensing in rare cases (e.g., Morgan et al. 2010). Most early disk RM experiments yielded inconclusive results. For instance, an early campaign on NGC  4151 built around the International Ultraviolet Explorer found a hint of the shorter-wavelength UV leading longer wavelengths, though the measured lag was not significantly different from zero (Crenshaw et al. 1996; Edelson et al. 1996). Further efforts to implement disk RM by correlating X-ray light curves gathered with space-based observatories with optical light curves typically from ground-based observatories (e.g., Wanders et al. 1997; Collier et al. 1998; Nandra et al. 1998; Collier et al. 1999; Collier et al. 2001; Suganuma et al. 2006; Arévalo et al. 2008; Arévalo et al. 2009; Breedt et al. 2009; Breedt et al. 2010; Cameron et al. 2012; Gliozzi et al. 2013) have often yielded suggestions of interband lags in the expected direction, but the results were never statistically significant ($>3\sigma$). Likewise, ground-based optical monitoring also yielded indications that the shorter wavelengths led the longer wavelengths (e.g., Sergeev et al. 2005; Cackett et al. 2007), but again not at a statistically significant level.

Recent observations have been able to produce more solid results by taking advantage of the unique capabilities of the Swift satellite, in particular, its ability to sample at high cadence across the X-ray/UV/optical regime needed to perform this experiment. Shappee et al. (2014) and McHardy et al. (2014) find clear evidence of the UV leading the optical in NGC 2617 and NGC 5548, respectively. The clearest previous measurement of interband lags was seen in a very large ($\sim$300 observations) Swift/HST/ground-based campaign on NGC 5548 (Edelson et al. 2015; Fausnaugh et al. 2016). A recent archival Swift survey by Buisson et al. (2016) reports evidence that X-ray variations lead the UV in several AGN, but also shows that detailed disk RM requires long-duration, high-cadence campaigns with multi-filter Swift UltraViolet/Optical Telescope (UVOT; Roming et al. 2005) data similar to what was done for NGC 5548.

This paper reports the results of intensive Swift monitoring of NGC  4151, with particularly detailed coverage in the X-ray regime, allowing us to measure temporal correlations and time lags between bands spanning an unprecedented wavelength range out to 50 keV. These results contradict the standard reprocessing model because the X-ray/UV lags are observed to be much longer than those within the UV/optical. This indicates that the arrangement of the emission components cannot be as simple as an X-ray corona that directly illuminates and drives a UV/optical-emitting accretion disk. Instead, the interband lags are consistent with the picture proposed by Gardner & Done (2017), which posits the existence of an energetically important emission component that peaks in the unobservable extreme-ultraviolet (EUV). While the peak of this putative component in the EUV cannot be directly observed, the “soft excess” seen in the X-rays and the “big blue bump” seen in the UV/optical could be interpreted as its high/low-frequency tails. The observed interband lags appear to be consistent with such an EUV component acting as an additional reprocessor that is illuminated and heated by the X-ray corona and then in turn illuminates and drives the variability in the accretion disk.

This paper is organized as follows. Section 2 summarizes the observations and data reduction, Section 3 presents the timing analysis, Section 4 discusses the challenges these results present for the standard reprocessing model and how the addition of a EUV component may solve these problems, and Section 5 gives some brief concluding remarks.

2. Observations and data reduction

2.1. Target

The target of this experiment, the Seyfert 1.5 galaxy NGC  4151 (redshift $z=0.00332$, de Vaucouleurs et al. 1991; distance $D\approx 19$ Mpc; Hoenig et al. 2014), is typically the brightest Seyfert 1 galaxy in the sky in the X-ray/UV/optical wavelength range accessible to Swift. For instance, the Swift Burst Alert Telescope (BAT) catalog (Krimm et al., 2013) indicates that NGC  4151 is twice as bright in the 15–50 keV band as the next-brightest Type 1 AGN. NGC  4151 is one of the Seyfert 1 galaxies in the original identification paper on these objects (Seyfert, 1943) and is often considered to be an archetype of the class (Ulrich, 2000). It is well known to be strongly variable across the wavelength range accessible to Swift (Edelson et al., 1996), making it an ideal monitoring target.

The bolometric luminosity of NGC  4151 is $L_{\rm bol}\approx 5\times 10^{43}$ erg s*-1* (Woo & Urry, 2002). The central black hole mass has been measured by RM to be $M_{\rm BH}\approx 43.57^{+0.45}_{-0.37}\times 10^{7}\,{\rm M}_{\odot}$ (Bentz et al. 2006, updated with the calibration of Grier et al. 2013), by gas dynamics to be $3.0^{+0.75}_{-2.2}\times 10^{7}\,{\rm M}_{\odot}$ (Hicks & Malkan, 2008), and by stellar dynamics to be $(3.76\pm 1.11)\times 10^{7}\,{\rm M}_{\odot}$ (Onken et al., 2014). NGC  4151 has been particularly well-studied in the X-rays. The soft X-rays are only weakly variable because that band is dominated by extended line emission (e.g., Zdziarski et al. 2002), but at higher energies (above $\sim$2 keV) the flux is strongly variable and thought to be coming from the corona. NuSTAR/Suzaku spectroscopy is consistent with reflection from the inner disk in NGC  4151 (Keck et al., 2015). X-ray time lags also show Fe K$\alpha$ reverberation in this object that would require reflection from the inner disk (Zoghbi et al. 2012; Cackett et al. 2014).

2.2. Observations

During 2016 February 20 through April 29, Swift executed an intensive monitoring campaign on NGC  4151 consisting of 319 separate visits of at least 120 s, an average of nearly 5 visits per day. These observations are summarized in Table 1. Start and stop times for Swift observations are originally recorded in Mission Elapsed Time (seconds since the start of 2001) and corrected for the drift of the on-board Swift clock and leap-seconds. These times were converted to Modified Julian Date (MJD), the standard for this observing campaign.

Swift observations with the UVOT were made in mode 0x037a, which allows for hardware windowing in the four longest-wavelength bands. This was done because this source is too bright to be observed in a standard, non-windowed mode. Observations with the X-Ray Telescope (XRT; Burrows et al. 2005) were made in Photon Counting (PC) mode, except for the last seven, which were made in Windowed Timing (WT) mode (Hill et al., 2004). The impacts of these observing modes on the data quality and other details are discussed in the following subsections.

These Swift observations were coordinated with intensive monitoring with numerous ground-based telescopes including the Las Cumbres Observatory Global Telescope (LCOGT) network and the Liverpool Telescope at La Palma. Those data will be presented in subsequent papers (K. Horne et al. in preparation; M. Goad et al. in preparation).

2.3. UVOT Data Reduction

six-filter, blue-weighted mode uvw1, u, b, and v) ″$\times$ 5″ indows. he frame time from 11 to 3.6 ms, he effect of pile-up (coincidence losses) in this bright source. The four hardware window observations are preceded by short (10 s) full-field exposures, his mode All UVOT data were reprocessed for uniformity, applying standard FTOOLS utilities (Blackburn 1995; from version 6.19 of HEASOFT111http://heasarc.gsfc.nasa.gov/ftools/). The astrometry of each field was refined using the AGN and up to 25 isolated field stars drawn from the HST GSC 2.3.2 (Lasker et al., 2008) and Tycho-2 (Høg et al., 2000) catalogs, yielding residual offsets that were typically $\sim$0.3″. Fluxes were measured using a 5″ ircular aperture, and concentric 40–90″ nnuli were used to measure the sky background level. The final values include corrections for aperture losses, coincidence losses, large-scale variations in the detector sensitivity across the image plane, and declining sensitivity of the instrument over time. After reprocessing, ere screened out to eliminate observations affected by tracking errors or with exposure times shorter than 20 s.

We use a non-default setting owever, when measuring  and the field stars with the highest signal-to-noise ratios (S/Ns), we found the resulting error estimates to be inconsistent with Gaussian statistics. This result is not surprising given that the UVOTAPERCORR documentation s not well established. We empirically examined a range of nd found that halving this parameter (to 7.5) yielded distributions much more consistent with Gaussian. For instance, 8.7% of the uvm2 and 76.8% of the uvw2 measurements fall within $\pm 1\sigma$ of the mean hese percentages are 72.8% and 64.1% when FWHMSIG = 7.5. ilter-dependent systematic errors of 0.92–1.08%.

The resulting light curves, shown in Figure 1, exhibited occasional anomalously low points (“dropouts”), especially in the UV. Similar dropouts were seen in an earlier Swift study of NGC 5548 (Edelson et al., 2015) and ound to be clustered in the detector plane. This may be due to localized regions of reduced sensitivity (Breeveld et al., 2016; it should be noted that the deviant flux points s such, we filtered discrepant points in the NGC  4151 data in a fashion similar to that in the Appendix of Edelson et al. This filtering consists of four steps: (1) identify dropouts from the light curves; (2) map the data onto the detector plane; (3) define boxes and (4) use ll data in these regions in the four shortest-wavelength bands. quadratic expression to each light curve in a sliding window of $\pm$2 days.

he number found in each band is given in Table 2.3.