Winds of Change: XRISM Resolve X-ray spectroscopy of NGC 4051

Abstract

NGC 4051 is a nearby (16.7 Mpc), Narrow Line Seyfert 1 galaxy (NLS1), which has a low black hole mass of $10^6$ M$_{\odot}$. It is also known for its rapid X-ray variability, on timescales of kilo-seconds and has a complex, multi component wind in both the soft X-ray and Fe K bands. Here we present the first high resolution XRISM Resolve spectrum of NGC 4051, which was captured in a historically bright state for a 150 ks exposure. XRISM resolves two blue-shifted Fe K shell absorption troughs in the mean spectrum, which can be ascribed to H-like iron and arises from two outflow components with outflow velocities of 0.025c and 0.04c. A time dependent spectral analysis shows that the iron K absorption is variable on timescales of less than a day, increasing in velocity over the duration of the observation. The velocity changes may be explained either by the passage of two separate transiting absorbers, of different velocities, or by a single accelerating outflow of approximately constant column density. In the latter case, the wind acceleration is likely to be too large to be caused by radiation pressure and instead magnetic driving is favored to accelerate the wind up to 0.04c. The outflow can originate from an accretion disk wind, whose kinetic power is sub-Eddington in contrast to recent examples of winds from powerful, luminous quasars observed by XRISM.

Figures (6)

• Figure 1: The X-ray variability of NGC 4051. The upper left panels show the 2025 light-curves (top to bottom) from Xtend, Resolve, EPIC-pn and NuSTAR. Time is relative to the start time of the XRISM Resolve exposure. The right-hand plot shows a zoom in of the 0.4-10 keV Xtend lightcurve, with finer time binning of 128 s per bin. Multiple short timescale kilo-second flares are present. The vertical dashed lines denote the time intervals used for the time sliced spectral analysis. The lower panel highlights the long term spectral variability, where the 2025 XMM-Newton EPIC-pn spectrum (bold black points) is observed at a historical high flux compared to the 15 EPIC-pn spectra taken in 2009.
• Figure 2: Time averaged spectra of NGC 4051 from the 2025 campaign. The top panel shows the XRISM Resolve spectrum (black), with the best-fit continuum model plotted as a red dotted line and the background (NXB) spectrum is shown in blue. The position of the strong Fe K$\alpha$ line and two absorption troughs are marked in red. The lower left plot shows the broad-band spectrum fitted with the model of Table 2. Here, Resolve is shown in black (grey errors) and NuSTAR FPMA in red. The individual model components (dotted lines) are as marked. The lower-right panel shows the residuals of the data to the model (in $\sigma$ units), but without including any outflowing absorbers. The two absorption troughs are apparent in Resolve and appear as a single trough in the lower resolution spectra. Here, EPIC-pn (sequence 2) is also overlaid in blue and XRISM Xtend in green.
• Figure 3: Time-dependent analysis of the Resolve spectrum, showing the mean spectrum (panel a) and the 5 slices (panels b--f) against the best fit absorber model as described in Table 3. The vertical dashed green lines mark the centroid energies of the two Fe K absorption troughs as measured in the mean spectrum at 7.14 keV and 7.26 keV respectively. In slices 1--3 the centroid energy of the absorption trough is consistent with the low energy trough at 7.14 keV, while in slices 4 and 5 the absorption line appears more blue-shifted coincident with the 7.26 keV trough -- see Table 3 for values. This may suggest an increase in velocity in the Fe K absorber during the observation. Note slice 5 also shows a higher absorption trough at 7.96 keV, which might originate from an ultra fast outflow at a velocity of $-0.13c.$
• Figure 4: Results from fitting the Resolve slices. The left plot shows the results from fitting a single absorption zone, allowing the outflow velocity to vary. The top panel shows the 2--10 keV luminosity, the middle panel the column density and the lower panel the outflow velocity. The absorber velocity increases through the observation, with a sharp jump occurring between slices 3 and 4. The right plot shows the case of fitting two separate absorbers to the slices, where the velocity is restricted to $v/c=-0.025\pm0.005$ and $v/c=-0.040\pm0.005$ for each zone. Here, changes in the column density of the two zones accounts for the variability. Thus the slower zone is apparent during the first 3 slices and is not detected in slices 4 and 5. Conversely, the faster absorber is only apparent in slices 4 and 5, with upper-limits only during slices 1--3.
• Figure 5: Results from fitting the SED. The left plot shows the fit to the data, where EPIC-pn is black, NuSTAR is red and the OM (U-band) point is a blue circle. The model fitted is the simple broken-powerlaw form of the SED, as described in the text. The right panel shows the comparison between the broken-powerlaw SED model (red dotted points), versus a physically motivated model computed with agnsed (black). In the latter, the bolometric luminosity is $L_{\rm bol}=3\times10^{43}$ erg s$^{-1}$. The models are corrected for reddening and line-of-sight absorption.