Estimating the masses of Narrow line Seyfert 1 galaxies using damped random walk method

TL;DR

This study recalibrates black hole masses for 1,141 NLSy1 and 1,143 BLSy1 galaxies using damping random walk (DRW) modeling of ZTF g-band light curves. It demonstrates that DRW masses computed with a constant $R_{ ext{Edd}}$ overestimate NLSy1 masses, and implements a multivariate calibration that explicitly includes $R_{ ext{Edd}}$ and rest-frame wavelength, with $R_{ ext{Edd}}$ estimated via Fe II strength and H$\beta$ line shape. Applying three calibration schemes, the authors consistently find that NLSy1s host lower-mass black holes than BLSy1s when accretion-rate effects are accounted for, though zero-point offsets remain relative to SE masses. The work also shows that forced ZFPS photometry improves DRW constraints and highlights the need for reverberation-mapped anchoring to establish robust DRW mass scalings for AGN demographics in upcoming time-domain surveys.

Abstract

Narrow-line Seyfert 1 galaxies (NLSy1s) are a subclass of active galactic nuclei (AGNs), commonly associated with rapidly accreting, relatively low-mass black holes ($10^6$ - $10^8 M_\odot$) hosted in spiral galaxies. Although typically considered to have high Eddington ratios, recent observations, particularly of $γ$-ray-emitting NLSy1s, have raised questions about their true black hole masses, with some estimates approaching those of Broad-line Seyfert 1 (BLSy1) systems. In this work, we present the recalibrated mass estimations for a large sample of NLSy1s galaxies with z $<0.8$. We apply the damped random walk (DRW) formalism to a comparison set of 1,141 NLSy1 and 1,143 BLSy1 galaxies, matched in redshift and bolometric luminosity using SDSS DR17 spectroscopy. Our analysis employs a multivariate calibration that incorporates both the Eddington ratio and the rest-frame wavelength to refine the mass estimates. We obtain median DRW-based black hole masses of $\text{log}(M_{\text{BH}}^{\text{DRW}}/M_\odot) = 6.25 \pm 0.65$ for NLSy1s and $7.07 \pm 0.67$ for BLSy1s, in agreement with their respective virial mass distributions. Furthermore, we identify strong inverse trends between the variability amplitude and both optical luminosity and FeII emission strength, consistent with a scenario where higher accretion rates suppress long-term optical variability. These findings reinforce the view that NLSy1s harbor smaller black holes and highlight the value of variability-based approaches in tracing AGN accretion properties.

Figures (11)

• Figure 1: Distribution of the refined sample of NLSy1s (Red) and their matched BLSy1 counterparts (Black) in the luminosity–redshift (L–$z$) plane. All sources have SDSS $r$-band magnitudes $\leq 19$ and were selected from the catalog of 2024MNRAS.527.7055P. The BLSy1 sample was matched using a maximum distance of 0.2 in the L–$z$ plane.
• Figure 2: Distribution of the number of epochs and temporal baselines for the ZTF $g$-band light curves. The left panel shows the archival (normal) photometry, while the right panel shows the forced photometry. Red solid-line histograms represent NLSy1 galaxies (final sample: 1,254), and black dashed-line histograms represent BLSy1 galaxies (final sample: 1,270). For the archival light curves, the median baseline is $\sim$2,000 days with a median of 580 high-quality epochs. In comparison, the forced photometry offers an extended baseline of $\sim$2,250 days and a higher median of $\sim$665 epochs, enabling improved sampling for variability studies.
• Figure 3: DRW modeling of NLS1 SDSS J000144.85+110959.9Left panel: The g-band light curve of SDSS J000144.85+110959.9 ($M_{\rm BH} = 10^{7.703}\,M_{\odot}$) DRW model is fitted with 1$\sigma$ uncertainty as orange shaded area. Right panel: Power spectral density (PSD) normalized and binned, shown with $1\sigma$ errors. The best-fit broken power law is overplotted in red and the DRW-based posterior prediction of the PSD is represented by the orange shaded band. The corresponding break frequency $f_{br}$ (from the broken power law fit) and $1/(2 \pi \tau_{DRW})$ (from the DRW fitting). The red shaded regions correspond to timescales greater than 20$\%$ the light curve length (in panels left lower and right lower) and less than the mean cadence (panel C), where the PSD is not well sampled.
• Figure 4: Comparison of the black hole mass offset, defined as $\Delta \log M_{\mathrm{BH}} = \log M_{\mathrm{BH, Zhou}}^{\text{DRW}}$(equation \ref{['zhou']}) - $\log M_{\mathrm{BH}}^{\text{SE}}$, as a function of the DRW timescale $\tau_{\mathrm{DRW}}$, for forced and normal BLSy1 light curves. The associated error bars correspond to the median absolute deviation scaled by $\sqrt{N}$ in each bin. The solid black and red lines show the best-fit MCMC regression models for the forced and normal samples, respectively, with shaded regions indicating the $\pm3\sigma$ confidence intervals. The vertical red dashed line marks the intersection point of the two fits at $\tau_{\mathrm{DRW}} \approx 72.3$ days, suggesting a potential threshold in timescale-dependent mass discrepancies between the two sampling regimes.
• Figure 5: The DRW damping timescale ($\tau_{\text{DRW}}$) as a function of single-epoch black hole mass estimates ($M_{\text{BH}}$) for the final sample of NLSy1s (red) and BLSy1s (black) galaxies. Right: The corresponding distributions of $\tau_{\text{DRW}}$, with solid red and dashed black lines for NLSy1s and BLSy1s, respectively. Upper: The corresponding distributions are single epoch mass $M_{\text{BH}}$, for NLSy1s (red) and BLSy1s (black).