The unusually red delay spectrum of the low-mass black hole AGN NGC\,4051 as revealed by intensive continuum reverberation mapping with the Las Cumbres Observatory

Abstract

We present a two-year optical reverberation mapping campaign of NGC 4051, an active galactic nucleus (AGN) hosting a low-mass black hole ($8\times10^5 M_\odot$), using daily observations in seven photometric bands from Las Cumbres Observatory augmented by archival data from Swift XRT and UVOT. The light curves show correlated variability with wavelength-dependent lags broadly consistent with the standard accretion disc scaling, $τ\propto λ^{4/3}$, and a pronounced u-band excess. However, the $i$ and $z_s$ lags are significantly larger than expected and cannot be explained by a combination of disc emission and diffuse continuum (DC) from the broad-line region (BLR), making NGC 4051 a notable lag-luminosity outlier. The spectral energy distribution (SED) of the variable AGN component is markedly redder than the canonical accretion disc prediction, $F_ν\propto ν^{1/3}$, typically observed in more massive systems. We explore two scenarios to account for the red UV-optical SED and the anomalously large $i$ and $z_s$ lags: (a) SMC-like dust reddening ($E(B-V)\sim0.18$) combined with optically thick emission from the inner edge of the dusty torus; and (b) a dominant diffuse continuum contribution. We discuss the implications of each scenario within a comprehensive multi-wavelength framework.

Figures (18)

• Figure 1: PyROA fit to 7-band LCO light curves spanning 2 years. The ROA lightcurve shape $X(t)$ has a Gaussian smoothing width $\Delta=1.5$ d. Each light curve panel shows data (colour-coded error bars) and the best-fit model light curve (solid black curve) with its $1\sigma$ uncertainty envelope (grey band), which expands greatly during the seasonal gap. All fluxes are in mJy. The subpanel below each light curve shows the normalized residuals ($-5<\chi<+5)$ to the jointly fitted PyROA model. Note that systematic trends remain in the residuals for the $r$, $i$ and $z_s$ bands. The right panel shows the marginalized posterior distribution for the time lag $\tau$, relative to the $g$ band light curve, with the median value and 68% confidence interval indicated by vertical solid and dashed black lines, respectively. Time delays are in observed frame days. The measured lags in each band are summarised in Table \ref{['tab:timedelays-two_years_sim']}.
• Figure 2: As in Fig. \ref{['fig:2year_fit']} except that the LCO light curves are de-trended by subtracting the ROA model with $\Delta=10$ d, (red dashed curves) and the ROA model with $\Delta=1.5$ d (black line) is fitted to the residuals. Note that systematic trends in the residuals are now largely eliminated. The resulting lags are also smaller by a factor $\sim2$, indicating that variations removed by the 10 d detrending are delayed relative to those captured by the $\Delta=1.5$ d ROA model.
• Figure 3: Top: The lag spectrum measured by PyROA with different smoothing widths $\Delta$, showing the general trend for $\tau$ increasing with $\Delta$ at all wavelengths, but particularly in the $u$ band and the redder $r$, $i$, and $z$ bands. This pattern suggests a 2-component model with rapid response from the compact accretion disc plus free-bound continuum response from the more extended BLR. Bottom: The faint (marked as stars) and bright (marked as circles) disc SEDs measured from the mean and rms of light curves measured from PyROA fits with different smoothing widths $\Delta$. The spectral shape is independent of $\Delta$ and suggests bound-free emission with Balmer continuum sampled by the $u$-band and Paschen continuum sampled by the redder bands. The absolute scaling of the inferred bright disc SED is independent of $\Delta$ ($\Delta$ colors are the same as the top panel). The faint disc SED is lower for $\Delta=0.65$ d and higher for $\Delta=20$ d.
• Figure 4: MEMecho fit to 7-band LCO light curves of NGC 4051. The observed fluxes (green points with error bars) are fitted with a linearised echo model with slow background variations, Eq. (\ref{['eq:memecho']}). The driving light curve (bottom panel) is convolved with 7 delay maps (left panels) to produce 7 echo light curves (right panels) plus slowly-varying backgrounds (red curves). MEMEcho adjusts the predicted light curves and delay maps (blue curves) to achieve a good fit ($\chi^2/N=1)$ to each of the light curves, while maximising the entropy to keep model curves positive and as smooth as possible. The delay map panels show the cumulative delay distributions (rising red curves). The median and quartiles of each delay map are marked by vertical red solid and dashed lines. These are listed both on the plot and in Table \ref{['tab:timedelays-two_years_sim']}. The delay maps all have a prompt component (disc) and the redder bands ($r$, $i$, $z_s$) have an extended response tail (BLR) and slow variations (torus). We interpret this as due to the BLR and the obscuring torus, respectively.
• Figure 5: Results of the flux-flux analysis. The colours have the same meaning as in Fig. \ref{['fig:2year_fit']}. The left panel shows the observed fluxes as a function of the dimensionless light curve shape $X(t)$ (from PyROA fit), which represents the relative variability of the driving continuum normalized to its mean level. The linear models fitted to the flux data adequately describe the variations in all bands, with the slopes $dF_\nu/dX$ given the RMS of the light curve at each wavelength. Extrapolating to fainter levels, the W2 flux is 1$\sigma$ above 0 at $X_{\rm g}$ (red dashed line). Evaluating the model fluxes at $X_{\rm g}$ gives the constant (host galaxy) SED. Subtracting this leaves the variable (AGN disc) SED, ranging from $X_{\rm F}$ at the faintest to $X_{\rm B}$ at the brightest during the campaign. The right panel shows the SEDs obtained from the flux-flux plot, corrected for Galactic dust extinction and reddening; $E(B-V) =0.011$SchlaflyFinkbeiner2011. The variable disc SEDs, faint, bright, and RMS, are marked in black and grey. The host galaxy SED (red) is much redder. The host galaxy flux inferred from the flux–flux analysis is consistent with the HST measurement of Bentz2013 ($f_{\nu,5100\text{\AA}} \approx 7.6$ mJy), providing an independent confirmation of the constant component.