WorksPhysics and Astronomy
Discovery of Ultra Fast Outflows in the Narrow Line Seyfert 1 Galaxy Mrk 110 with XMM-Newton and NuSTAR
Abstract
We present an X-ray analysis of XMM-Newton and NuSTAR spectra of the Narrow-Line Seyfert 1 (NLS1) galaxy Mrk 110. Our analysis reveals the intriguing presence of a multi-component ultra fast outflow (UFO) both in the grating spectrum and the NuSTAR data. The moderate UFO significance (3-4), in the grating and CCD data indicates a shallow absorber, consistently with the absence of previous such claims in this source. The two components (UFO1 and UFO2) in the grating spectrum show a fairly wide range of velocities (0.05-0.06c), as their respective ionization state (Log U = -1.43 and -1.82), however they share the same column density (Log N 19.7). In the NuSTAR spectrum, we found a much faster wind component with a velocity of 0.21c (UFO3). We show that, in an energy-driven wind scenario, UFO3 might carry enough energy to produce significant feedback on its host galaxy. Considering its known radio properties, Mrk 110 is therefore another Narrow-Line Seyfert 1 with stratified ultra-fast outflows co-existent with a pc-scale jet/outflow radio structure.
INTRODUCTION
Active Galactic Nuclei (AGN) are very luminous compact regions observed at the center of galaxies, whose extremely high luminosity output is not directly attributed to stars. They are powered by supermassive black holes (SMBH) that are actively accreting matter, which generates an enormous amount of energy [56]. Among their many properties, active nuclei are characterised by multi-scale outflows observed in various bands of the electromagnetic spectrum. [22, 23].
Ultra-fast outflows (UFOs) are high-velocity, massive and powerful winds of gas that are launched from the central regions of AGN. These outflows have velocities larger than km s. They were detected in the X-ray spectra of AGN and were usually observed in the form of absorption lines of ionized gas where the only dominant ions are helium-like and hydrogen-like Iron (e.g. Tombesi et al. [62]). Several studies have documented the presence of UFOs in specific AGNs ([9, 54, 34] (Longinotti et al. 2015)) and according to systematic studies, outflows have been observed in 30-40% of nearby AGN ((Matzue et al. (2023)) and reference therein). Tombesi et al. [62] and Gofford et al. [17] suggest that UFOs are likely launched from the accretion disk scales. The mass outflow rate of these winds (–) and their kinetic energy (–) are consistent with the predictions made by theoretical models of black hole feedback [29].
AGN feedback, which refers to the impact of AGN activity on the large scale environment, is thought to have a profound effect on the evolution of the host galaxy properties [26]. Simulations reproducing the galaxy luminosity function at high luminosities require a mechanism to regulate the star formation rate, and to produce the observed relationships between SMBH and their host galaxies [10, 6]. AGN winds have been proposed as one of the mechanism behind Quasar feedback [32]. At present, the relation of the inner X-ray wind to galaxy-scale outflows, supports the idea that fast winds launched close to the AGN accretion disk may act as a trigger for altering star formation in the host galaxy and start feedback processes [13, 61, 37, 39, 36].
NLS1 galaxies form a subclass of AGNs characterized by the presence of an H line with a FWHM and an unusually prominent Fe II spectral lines [47]. Brandt et al. [7], Boller et al. [4], showed that NLS1 exhibit X-ray properties different from the general population of Seyfert 1 galaxies. They have steep soft- and hard-X-ray spectra and show fast and large amplitude variability. These properties suggested that NLS1 have relatively smaller BH masses and higher Eddington accretion rates. As such NLS1 occupy one extreme end of the AGN "eigen-vector 1" [5]; see also [63]. It has been shown that NLS1 have super-solar metallicity [14 b] and that they likely undergo secular growth of their BH [18, 42 a]. Their host galaxies have pseudobulges, further supporting the interpretation of secular black holes growth [41]. NLS1s may represent an early evolutionary stage of AGNs [40].
Recent studies have uncovered UFOs in soft X-ray X-ray high-resolution spectra of NLS1 [21, 20, 49, 57, 34] (Longinotti et al. 2015; Xu et al. 2024). High-resolution spectroscopy has revealed that soft X-ray UFOs can consist of multiple components with distinct velocities, ionization states, and column densities, and that, in some cases, they can be massive enough to produce feedback, as observed, for example, IRAS 17020 + 4544 [36] (Longinotti et al. 2015).
Mrk 110 is a NLS1 at with an X-ray luminosity of in the 2-10 keV band. Assuming a bolometric correction (Duras et al., 2020), the bolometric luminosity for this source is . It was found variable in X-rays as well as in optical/UV wavelengths [19, 31]. Kollatschny [30] estimated a SMBH mass of via reverberation mapping methods. Given the Eddington luminosity of (see 4.3) this gives a high accretion rate of .
As to its X-ray history, Boller et al. [3] reported the presence of broad and narrow emission lines in the XMM-Newton Reflection Grating Spectrometer (RGS) spectrum, finding that the distance from the SMBH to the emission line region ranges from 0.2 to 1 light day from the SMBH. Reeves et al. [59] reported in the RGS spectrum a broad component of the OVII emission line with a FWHM of , that was modeled with a relativistic profile from gas in the inner accretion disc. Based on simultaneous observations of XMM-Newton and NuSTAR, Porquet et al. [53] showed that the dominant physical processes above 3 keV are hard Comptonisation in a hot corona and soft Comptonisation in a warm, optically thick corona. In this scenario, the warm corona accounts for the soft X-ray excess, while the hot corona produces the primary power-law continuum. According to these authors, the latter with mildly relativistic disc reflection are responsible for the broad O VII and Fe K emission lines. Regarding other observing bands, Panessa et al. [48] showed evidence for variability at 5 GHz on scales of days to weeks time-scales and suggested that emitting regions have a size smaller than Schwarzschild radii. Considering the high brightness temperature and the radio and X-ray variability, these authors proposed that synchrotron emission from a low-power jet and/or an outflowing corona is favored to explain the radio and X-ray timing properties.
More recently, Juranová et al. [27] report the presence of broad X-ray lines in the grating spectrum that can be accounted for by primarily UV-emitting gas in their model. Xiang et al. [65], using XRISM observations of NGC 4151, revealed a stratified, multiphase accretion-disk wind in which warm absorbers, fast outflows, and UFO coexist, with mass outflow rates comparable to or exceeding the accretion rate. Notably, two UFO components carry sufficient kinetic power to produce significant AGN feedback, broadly consistent with magnetocentrifugal driving. Mizumoto et al. [46], using XRISM observations of PG 1211+143, resolved the Fe K absorption features into six distinct UFO velocity components ( to ), revealing a complex UFO forest structure. The inferred mass outflow rate () is comparable to the Eddington accretion rate, indicating that disk winds constitute a major channel of mass and energy release in near-Eddington AGN.
Observations and Reduction
XMM-Newton
The XMM-Newton satellite observed Mrk 110 six times between November 2004 and April 2020 (see Table 1). Five of these datasets are used in this study. The 2004 observation is not considered in this work due to the large temporal gap with respect to the other five observations, which were obtained within less than five months of each other. The data were processed using the rgsproc task of the software System Analysis Software (SAS) 21.0. The spectra of the rgs1 and rgs2 cameras were combined with the RGSCOMBINE task. To improve the Signal-to-Noise Ratio (SNR (S/N = counts/sqrt(counts)) and given the proximity in time of the observations and their nearly constant flux, the four spectra from 2019 and the one from 2020 were combined into one spectrum with the RGSCOMBINE task. The total exposure accumulated in this combined spectrum is 220.2 ks. The EPIC-pn exposures were processed by applying the standard filtering criteria through the SAS task epproc. The source spectra were extracted from circular regions of 40″. Background regions were selected in source-free portions of the CCD. Response matrices were provided by the SAS tools arfgen and rmfgen.
| XMM-Newton Obs ID (RGS) | Obs Date | Exp ks | Flux (RGS) 0.3-2.0 keV |
| 0201130501 | 2004/11/03 | 47.4 | |
| 0840220701 | 2019/11/03 | 43.6 | |
| 0840220801 | 2019/11/05 | 43.0 | |
| 0840220901 | 2019/11/07 | 40.6 | |
| 0852590101 | 2019/11/17 | 44.5 | |
| 0852590201 | 2020/04/06 | 48.5 | |
| NuSTAR | 3.0-15.0 keV | ||
| 60201025002 | 2017/01/23 | 184.5 | |
| 60502022002 | 2019/11/16 | 86.7 | |
| 60502022004 | 2020/04/05 | 86.7 |
Table 1. Log of the XMM-Newton RGS and NuSTAR observations of Mrk 110.
NuSTAR
NuSTAR [24] observed Mrk 110 in January 2017, November 2019 and April 2020 (see Table 1), with its two coaligned X-ray telescopes with corresponding Focal Plane Module A (FPMA) and B (FPMB). The total elapsed time of each of the three pointings exposures (Obs IDs 60201025002, 60502022002 and 60502022004) is 386 ks, 163 ks and 168 ks, respectively. The level 1 data products were processed with the NuSTAR Data Analysis Software (NuSTARDAS) package (v. 2.1.2). Cleaned event files (level 2 data products) were produced and calibrated using standard filtering criteria with the NUPIPELINE task and the latest calibration files available in the NuSTAR calibration database (CALDB 20230321). Background spectra were extracted from source-free circular regions with a radius of 80 arcsec. Extraction radii for the spectra of the source were computed via an iterative process which leads to the maximization of the SNR in the 3–10 keV energy band, similar to the approach described in Piconcelli et al. [50]. We therefore adopted circular regions centered on the source with radii of 113 arcsec for the three FPMA spectra and 108 arcsec, 98 arcsec and 98 arcsec for the FPMB ones. FPMA spectra were binned in order not to over-sample the instrumental resolution more than a factor of 2.5 and to have a SNR greater than 5 in each spectral channel, the same energy binning was then applied to the FPMB spectra. The net observing times for the FPMA/FPMB spectra of the three data sets are 184.6 ks/184.3 ks, 86.8 ks/86.2 ks and 88.7 ks/87.9 ks, respectively.
Spectral Analysis
RGS Data
We carried out the spectral analysis of Mrk 110 RGS data using the unbinned spectrum, making use of the Cash statistics ([8]), The software used for the spectral analysis was XSPEC v.12.15.1 (X-Ray Spectral fitting Package). Fit parameters are quoted with error bars throughout the paper.
Modeling the Spectrum
We fit the X-ray continuum in the 7-35 Å (0.35 - 2.5 keV) interval with a power law affected by Galactic absorption of column density ([28]) using the model TBabs ([64]) of XSPEC. The photon index of the power law is , consistent with typical values obtained in the soft X-ray band for NLS1 ([16]). The model yields fit statistics of for 2797 degrees of freedom (d.o.f.). We found residuals hinting to the presence of emission and absorption features in the spectrum, therefore we included narrow components of the O VIII Ly and O VII triplet (, , ) emission lines. We also included the absorption lines by O VII at 21.601 associated with hot Galactic/CGM gas, and the NI absorption line at 31.3 Å associated with cold Galactic gas (Figure 1). The normalization of the NI absorption line is tied to that of the O I absorption line as such that their relative ionic column densities follow the ratio predicted by the TBabs
component (i.e. ). For , this gives and , hence . We therefore fixed , which yields . The position and width of the narrow emission and absorption lines were fixed and we only left normalization free to vary. The model with the four narrow emission lines and the two absorption lines results in , and this model provides an improvement to the simple power law fit of .
Following the recent findings by Juraňová et al. [27] on the presence of broad emission lines in the RGS spectrum, we added six broad emission lines as well as one Warm Absorber (WA) component. Although Juraňová et al. [27] used a second WA component with , the inclusion of this additional component in our model did not lead to any statistically significant improvement. Therefore, we opted to keep one single-component WA in our baseline model, fitted by the PHASE photoionization code (see details in § 3.1.2). The model including six broad emission lines and one warm absorber (WA) provides a fit statistic of . The addition of the six broad lines plus the WA improves the fit by . Table 2 lists the best-fit parameters of the six broad emission lines, while Table 3 shows the parameters of the WA. The WA parameters, including their uncertainties, are consistent with those of the first WA reported by Juraňová et al. [27].
| Obs. | Rest Wavelengths | Intensity | Sign | Line ID | Component |
| Emission | |||||
| 12.14 | 2.3 | Ne X | Broad | ||
| 13.85 | 2.5 | Ne IX | Broad | ||
| 18.96 | 7.5 | O VIII | Narrow | ||
| 21.6 | 4.9 | O VII | Narrow | ||
| 21.8 | 3.7 | O VII | Narrow | ||
| 22.101 | 8.4 | O VII | Narrow | ||
| 22.101 | 8.4 | O VII | Broad | ||
| 24.78 | 5.2 | N VII | Broad | ||
| 29.53 | 2.4 | N VI | Broad | ||
| 33.74 | 7.5 | C VI | Broad | ||
| Absorption | |||||
| 12.86 | 3.6 | Fe XX | ? | ||
| 22.02 | 2.0 | O VI | UFO1 | ||
| 22.37 | 4.0 | O V | UFO2 | ||
| 2.1 | ? | ? | |||
| 28.87 | 2.0 | N VI | UFO1 | ||
| 32.75 | 2.7 | N VI | UFO2 | ||
| 33.42 | 2.7 | C V | UFO2 | ||
| CGM | |||||
| 21.602 | 3.1 | O VII | |||
| 31.3 | 3.9 | N I |
Table 2. List of Emission and Absorption Lines Detected in the RGS data. Note. The identification of lines were made by matching this phenomenological fit to the physical model in Table 3.
| Component instrument | Log | Log (cm) | (km s) | Statistics | AIC | Factor |
| WA | 16/3 | 2943 | 148 | |||
| UFO1 | 21/3 | 2928 | 1808 | |||
| UFO2 | 16/3 | 2916 | 403 | |||
| 528 | ||||||
| UFO3 (NuSTAR) | 18/3 | 512 | 2981 |
Table 3. List of WA and Ultra-fast Wind Detected in the RGS and NuSTAR data of Mrk 110.
A visual inspection of the RGS spectrum fitted with this model still shows negative residuals throughout the entire spectral range. These suggest absorption lines by ionized gas, so we performed a blind search for individual absorption lines by adding one narrow Gaussian component with negative intensity to our model, letting its position and intensity vary freely over the whole spectral range. The line width was fixed to 0.1 Å. This procedure was repeated for as many lines as required until the line significance drops below . Table 2 reports the 7 absorption lines detected above a threshold with this procedure. The intensity of the line was divided by the positive error to estimate the line significance. Finally we note that the line at Å is identified with O I from hot gas in the Galaxy. Figure 1 presents our best-fit model with all the lines reported in Table 2 plus the WA listed in Table 3.

Figure 1. RGS spectrum in the 10-35 Å range, fitted with a model including a power-law continuum and eighteen Gaussian components (four narrow and six broad in emission and eight in absorption). The O VIII Lyα, O VII triplet (r, i, f) for narrow emission lines, the six broad emission and the absorption for O VII and O I of our Galaxy at are marked in magenta. The other absorption lines are reported in Table 2. The feature at 31.3 is a blend of CV from UFO and and NI from CGM. The identifications were obtained after performing the spectral modeling and fitting of the data. The spectrum was binned by up to three instrumental channels for display purposes only and are shown in the observed wavelength.
Although most absorption lines in Table 2 seem to have marginal significance, two of them do show significance larger than . The presence of multiple negative residuals suggest that they could be the result of a shallow absorber in the line of sight. However, the positions of the lines as reported in Table 2 do not align with any obvious transitions in the reference framework of the source, suggesting the presence of an UFO, where absorption lines are significantly shifted from their expected positions. To test for the presence of such outflowing gas, we proceeded to model the spectrum with a self-consistent ionized absorber, as described in the following.
Modeling the Ultra Fast Outflow
To achieve a physical description of the absorption in the RGS spectrum of Mrk 110, we use a self-consistent model, specifically the photo-ionization code PHASE (Krongold et al., [34]; Longinotti et al., 2015). The model assumes a slab of ionized gas in a plane-parallel geometry, and it is described by four parameters: the ionization parameter1, the equivalent hydrogen column density of the slab , the outflow velocity of the wind
and the micro-turbulence velocity of the medium. Given that the Gaussian absorption lines in the spectrum are narrow, we fix the micro-turbulence velocity to 10 .
The Significance of the UFO Components
While fitting the data with the PHASE components, we evaluated the robustness of a model including a UFO using two different statistical methods.
First, we used the Akaike Information Criterion values (Akaike, [1]), as done by Hebbar et al. [25]. These values are defined as:
where is the number of parameters and is the value of the −statistic for the best fit with a given model2. The likelihood that the model including the UFO component is statistically preferred over the model without it, is given by the quantity . Hebbar et al. [25] pointed out that the inverse of this quantity represents the factor by which the more complex model, which in our case contains the UFO, is preferred over the simpler model.
The second method to estimate the statistical significance of the UFO components consisted of Monte Carlo simulations. To accomplish this, we followed the procedure in Longinotti et al. (2015) (hereafter L15). We assumed the phenomenological continuum model described in § 3.1.1 (comprising a power law with Galactic absorption plus 10 emission lines and one Galactic absorption line) as our baseline model for the simulations. Then, we incorporated each identified UFO components into
the baseline model, as done in the actual spectral fitting. For instance, in the simulations, we included the baseline model to evaluate the significance of UFO1 and baseline+UFO1 to assess the significance of the UFO2. For each UFO component, we generated 1000 synthetic spectra using the XSPEC package, each with the same photon statistics as the real data set. We compared the distributions of between the model with and without the addition of the UFO component to the values measured in reality (). The fraction of simulations where is lower than tells us the confidence level, and this conclusion led to the significance values reported in the last column of Table 3.
Fitting the Ultra Fast Outflow
Initially, we fit the data with a model consisting of a power law continuum absorbed by a slab of photo-ionized gas modeled by a PHASE component in the 7–35 Å (0.35 - 2.5 keV) range, plus Galactic absorption, plus the broad and narrow emission lines, plus the CGM absorption line, and the WA described in § 3.1.1. The ionization parameter, the column density, and the outflow velocity of the absorber were left free to vary. The addition of one PHASE component improves the fit of d.o.f. ( for 2780 degrees of freedom), indicating the likely presence of ionized gas along the line of sight. The absorber (hereafter UFO1) has an outflow velocity of (see Table 3). Given its low level of ionization () and column density (), UFO1 produces absorption by ions such as O VIII, O VII, O VI, O V, C VI, C V, C IV and N VI (red labels in Figure 2). The significance of UFO1 was evaluated as described in § 3.1.3. Comparing AIC statistics suggests that the model including UFO1 is preferred by a factor of 1808 over the model that only considers the continuum. Using Monte Carlo simulations UFO1 have a statistical significance of . After the addition of UFO1, other residuals remain in the spectrum. Therefore, we continued fitting the data by iteratively adding additional PHASE component to the model (as we did with UFO1, for each component, the ionization parameter, column density, and outflow velocity were free parameters). One additional statistically significant PHASE component was found. This component is hereafter designated as UFO2. Table 3 presents the results of our best fit and Figure 2 shows the best fit of the RGS spectrum.

Figure 2. RGS spectrum in the 10-35 Å range, fitted with a model including a power-law continuum attenuated by one WA and 2 UFO components. The absorption lines produced by the WA are labeled in green, for UFO1 are labeled in red and those by UFO2 are marked in blue. The O VII and O I from the CGM of our Galaxy at are marked in magenta. The spectrum was binned by up to three instrumental channels for display purposes only and are shown in the observed wavelength. Absorption lines marked in green are associated with the WA component, while those shown in red and blue correspond to UFO1 and UFO2, respectively. These identifications were obtained after performing the spectral modeling and fitting of the data.
UFO2 has a slightly higher velocity than UFO1, its value is . UFO2 has low ionization (Log ) and column density of (Log ). The absorption lines associated to UFO2 are O VII, O VI, O V, O IV, N V and C V (blue labels in Figure 2). UFO2 improves the fit, with d.o.f. (). The AIC statistics indicates that UFO2 improve the fit by factor of and the Monte Carlo simulations from UFO2 shows a statistical significance of .
The final best-fit model includes the power-law continuum, plus four narrow and 6 broad emission lines, plus the CGM absorption line, plus one WA and the two UFOs components reported in Table 3. The final RGS best fit is shown in Figure 2.
Given that the absorption lines imprinted by these absorbers are shallow and difficult to recognize, we carried out further checks by allowing identification by other species. First, we tested if the O V line in UFO2 could be instead O IV or O VI in two different ways, i.e by changing the ionization parameter and changing the outflow velocity of the PHASE component. In both cases, the statistic got worse with the case of the line being O IV yielding an improvement of d.o.f. (); in the case of the line being identified as O VI, the improvement in the statistics was even smaller d.o.f. (). We tried the same test for UFO1 where O VII might be instead identified as O VI or O V. In both cases the model with the alternative identifications did not improved the statistic with respect to our proposed interpretation. In addition the absorption pattern corresponding to the alternative identifications result in a model that produces other absorption features clearly not consistent with the data. With this test we further corroborate that the identification of the lines is correct and the best fit reported Table 3 is the most consistent with the data.
Our fit still shows some negative residuals, particularly around 11.69 and 24.6 Å (see Table 2), where negative Gaussian profiles with significances of 3.6 and 2.1 were required. The line at 11.69 is consistent with absorption by Fe XX. The feature at 24.6 Å cannot be accounted for by adding another significant UFO component or by the presence of a Galactic line at , and therefore, it remains unidentified in our analysis. The line at 30.77 Å could be consistent with absorption by C V at the same velocity of UFO2, however, this component does not have enough neutral material to produce this feature.
As a final remark we noted the spectrum shows positive residuals around 23 to 23.5 Å. These residuals were modeled by Reeves et al. [59] with a relativistic line profile. Since they fall right next to the galactic line O I (23.5 Å) and none of the detected absorption features fall in this spectral portion, we decided not to include the modeling of these residuals in our analysis.
Searching for the highly ionized fast wind in CCD spectra
Motivated by the findings of the ultra-fast wind in the RGS data, we analysed the NuSTAR and XMM-Newton EPIC pn spectra to search for higher ionization absorption by Fe XXV or/and Fe XXVI.
The NuSTAR spectrum
We begin our analysis with NuSTAR since it has a broader bandwidth than the XMM-pn, so that we can later compare it with the XMM-Newton pn data. We used statistics, given the binning applied to the data. First, we simultaneously fit the NuSTAR spectra from the three observations by adding an energy-independent factor (model constant of XSPEC) to the continuum for small calibration uncertainties from the FPMA and FPMB detectors.
We fit the X-ray continuum in the interval of 3-15 keV, using a power law modified by Galactic absorption of column density of (Kalberla et al. [28]) with the model TBabs (Wilms et al. [64]) of XSPEC. The photon index of the power law is . This model provides an acceptable representation of the data, with a . However, we visually noted evidence of residuals pointing to the presence of an emission line from Fe I-XII K fluorescence at 6.330.02 keV. Adding a Gaussian profile to fit this line component improves the fit statistic to .
In addition to the Fe I-XII K fluorescence line, we found evidence for negative residuals in this spectral band. These negative residuals were fit by adding a narrow Gaussian line to the above model with the intensity allowed to vary between positive and negative normalization. The position of the line was also free to vary. The statistical improvement derived by the addition of the absorption feature was . The position of the line is keV where absorption features by Fe XXV or Fe XXVI from highly ionized gas at mildly relativistic velocities might be expected. Figure 3 shows the 68%, 90%, and 99% confidence contours between the absorption line energy centroid vs the line intensity, confirming a statistical significance larger than 3 . Assuming that absorption is due to Fe XXV or Fe XXVI, the position of the line would indicate the presence of absorbing gas with a velocity ranging from to , respectively.

Figure 3. Confidence contour plot between the absorption line centroid energy and flux (68% (blue), 90% (green), and 99% (red)), from the simultaneous fits for the three observations of NuSTAR data.
After carrying out the simultaneous analysis on the 3 NuSTAR observations we decided to look for the line in each observation separately. In the year 2017, the line was measured with a statistical significance of 2, in 2019 the significance reached a level above 3, while in 2020, a significance of 2 was observed. Figure 4 shows the model fitted to the NuSTAR data of 2019.

Figure 4. NuSTAR spectra of Mrk 110 in the 3-15 keV range, fitted with a model including a power-law continuum attenuated by Galactic absorption. One emission line produced by Fe I-XII K fluorescence and one absorption line interpreted as outflowing gas from Fe XXV or Fe XXVI are included in the model.
The joint NuSTAR-XMM-Newton EPIC-pn spectrum
We decided to support the analysis of NuSTAR spectrum by including in our analysis the EPIC-pn spectra from 2019 and 2020, therefore the ones contemporaneous to NuSTAR (see Table 1). We used the combined pn spectrum for these two epochs and produce a joint fit NuSTAR-XMM. The continuum baseline model of NuSTAR, consisting of a power-law and a Fe I-XII K fluorescence at 6.330.02 keV line, was applied to the joint CCD spectrum. The inclusion of the XMM-Newton pn spectra allowed us to find an additional emission feature at 6.67 keV, resulting in a significant improvement of for 2 d.o.f. This line is consistent with emission from FeXXV, as commonly found in several AGN. e.g. Longinotti et al. [35] Bianchi et al. [2] and Zaidouni et al. [67]. We also attempted to fit additional absorption lines at 6.97 keV and 7.06 keV respectively corresponding to FeXXVI and Fe I K transitions; however, these features led to only marginal improvements (), therefore they were not included in the final model. Once the two significant emission lines were included in the model, we search
for the absorption line in the joint fit and found it at 8.33 keV, resulting in an improvement of . The EPIC-pn spectrum does not show the absorption feature as clearly as the NuSTAR spectrum when inspected alone. However, in the join EPIC-pn/NuSTAR fit, the absorption line is recovered at keV, with an improvement of for two additional free parameters. This centroid energy is fully consistent with the value obtained from the NuSTAR-only fit, keV. We therefore conclude that the EPIC-pn data are consistent with the presence of the absorption feature, although they do not independently provide a stronger detection than NuSTAR.
Applying a physical model to NuSTAR spectrum
After establishing that the absorption feature is recovered in the joint EPIC-pn/NuSTAR fit with a centroid energy consistent with the NuSTAR-only analysis, we proceeded as in § 3.1.4 and applied a PHASE component to the NuSTAR data. Since the joint EPIC-pn/NuSTAR fit does not provide significantly different line parameters with respect to NuSTAR alone, and since the EPIC-pn spectrum does not add strong independent constraints on the absorber, we performed the physical modeling and the computationally expensive Monte Carlo simulations using the NuSTAR data only. Therefore, the significance quoted below for UFO3 refers to the NuSTAR-based PHASE analysis, while the EPIC-pn spectrum is used as an independent consistency check. Upon multiple iterations with distinct parameter sets and comparison to the transitions predicted by PHASE, we could determine that the absorption line is most likely identified as Fe XXVI. We, therefore, labeled this additional absorber UFO3. The associated parameter values for UFO3 are Log , Log and . The model improved the fit by . Given the high ionization state of UFO3, the expected absorption lines in the spectrum are Si XIV, S XVI, Fe XXIV, Fe XXV, and Fe XXVI. Unfortunately, none of these lines can be seen in the RGS data (they all fall below 10 Å), therefore a cross-check for this particular UFO component cannot be properly performed. Finally, the AIC statistics indicates that the inclusion of the UFO detected in the NuSTAR spectrum improves the fit by a factor of ≈ 2981. To assess its statistical significance, we performed Monte Carlo simulations following the procedure described in § 3.1.4. We generated 1000 synthetic
spectra based on a simultaneous fit to the 2017, 2019, and 2020 observations, adopting a baseline model consisting of a power-law plus two Gaussian lines at 6.33 keV and 6.67 keV. From this analysis, we obtained a statistical significance of ∼ 4σ for the UFO component.
Discussion
Despite the moderate significance of the absorbers found in our analysis, two UFO systems are detected at . Previous studies by Reeves et al. [59] and Porquet et al. [53] were primarily focused on emission originating from the disk and the warm and hot corona. This may explain why no absorption was reported. In the following, we discuss our interpretation of the shallow ultra fast outflow observed in Mrk 110.
Interplay of Light and Heavy UFOs
Mrk 110 is another case of a Narrow Line Seyfert 1 Galaxy with multiple UFO components. Here, three different UFO systems were detected. Two of these outflows (UFO1 and UFO2) have quite similar column densities (Log , see Table 3) and show the lower range of outflow velocity (∼26,000-29,000 km/s) measured in this analysis. We will refer to these two systems as "light" UFOs. Within the errors, UFO1 and UFO2 have remarkably similar ionization parameters. The third UFO (UFO3) reported in NuSTAR spectra, is clearly different from the two systems, with 2 to 3 orders of magnitude larger column density and ionization state. UFO3 is also the fastest, with an outflowing velocity larger by a factor of 2 than the light UFOs. We will refer to this system as a "heavy" UFO.
Mrk 110 exhibits different features compared to a simple wind produced by one or two lines of highly ionized Fe. This is in line with other recent studies where stratification in the flow was observed in NLSy1 galaxies ([34] (Longinotti et al. 2015)). As in those objects, light UFOs in this source are shallow and moderately ionized. One current view of stratified light UFOs is that they are associated with the shock of a heavy wind with the interstellar medium (ISM) of the host galaxy producing high-velocity entrained clumps of gas, and/or with related hydrodynamical instabilities in the flow ([55, 60]). The complexity found in Mrk 110 with the different velocity in the heavy component poses yet additional clues to this view. The presence of the heavy UFO in other sources has been interpreted in terms of perpendicular motion of the gas with respect to our line of sight ([52, 58]). The particular view offered by Mrk 110 suggests that both processes (shocks and perpendicular motions) are at play, with a scenario more complex than the simple radial flows explored in numerical and analytical models of BH-ISM interaction and feedback ([29]). An overall scenario seems to arise, where the wind consists of a multi-phase heavy flow interacting with the host galaxy ISM, resulting in the light components.
Narrow Line Seyfert 1 and its Relationship with Ultra-Fast Outflows
The detection of high velocity winds in Mrk 110, a NLS1 galaxy, adds to the growing number of sub-relativistic winds found in this type of AGN. The high accretion rates in NLS1 may play a role in the launching mechanism of these winds, which have been suggested to be radiatively driven ([51] (Matsue et al. 2017)). We note however that recent research has questioned the efficiency of this process in accelerating gas to subrelativistic speeds ([38]). Nevertheless, whatever the process responsible for launching the wind, a tight connection between high accretion and UFOs is clearly emerging, given the steady increase in the number of detections of ultra-fast outflows in NLS1 over the last decade ([20, 49, 33, 57, 34] (Longinotti et al. 2015)).
Another tantalizing piece of the puzzle relates to the UFO and radio emission connection, specifically in highly accreting AGN with fast X-ray outflows, where pc-scale radio jets have been discovered (see Longinotti et al. [36] and reference therein). Indeed, a recent VLBA-X-ray campaign on Mrk 110 (Panessa et al. [48]) revealed a variability pattern that among other alternatives, may be interpreted as due to a jetted or outflowing structure on sub-pc to pc scales. The present discovery of UFO processes in this source opens the door for further investigations on the coincidence of low power radio outflows/jets and X-ray winds in AGN.
Feedback implications
Using simple energetic considerations, we can estimate if UFOs have the capacity to carry enough energy to suppress star formation in their host galaxy and influence its evolution. In accordance with Hopkins & Elvis [26], a wind that transports energy outward in an energy-driven mode requires a minimum power level for feedback equivalent to 0.5% of the Eddington luminosity. As we illustrated below, UFO3 observed in Mrk 110 align with this criterion, suggesting their potential to satisfy this condition.
To estimate the energetics of the outflows in Mrk 110, we assume that the outflow velocity is larger than or equal to the escape velocity at the launch radius. We take this to be . Under this condition, the mass outflow rate for a radial wind in spherical geometry crossing our line of sight can be estimated by
where and correspond to the mass of the proton and to the mean atomic mass per particle (), to the covering fraction of the outflow, and and to the column density and the observed velocity of the outflow, respectively (Tombesi et al. [61]). Following Longinotti et al. (2015), we can express this formula in units of () and 10,000 km s () as
The energy outflow rate, or kinetic luminosity, is
We found that UFO3 is the only component that could produce feedback. Using , , and from Table 3 for UFO3 yields
and
The Hopkins & Elvis [26] criterion for triggering feedback by an energy-conserving wind for UFO3 is
for very conservative covering factors. If for UFO3 is this component is capable of significant feedback over their host galaxy. It is important to highlight that these estimates come with certain caveats related to the geometry of wind, which contributes to the computation of . Another source of uncertainty is the black hole mass determination, which goes into the estimation of . Considering this, the above ratio should be interpreted with caution.
Previous studies have shown that 30% to 40% of Seyfert galaxies show UFOs [62, 17]. If we consider this range of 30–40% as a proxy for the wind covering factor, the heavy UFO in Mrk 110 may satisfy the theoretical feedback condition for covering factors near the upper end of this range. An evolving perspective continues to grow: Winds in NLSy1 galaxies appear to have full potential for substantial AGN feedback in their host galaxies.
Conclusions
The presence of the UFOs shown herein is based on the simultaneous detection of three different UFOs components in XMM-Newton and NuSTAR. These UFOs are fitted with PHASE a self-consistent absorption model, and despite their moderate significance, their properties have been fully characterized and validated through Montecarlo methods, showing a combination of shallow absorbers with heavier components.
We show that under certain physical assumptions, the heavy UFO is capable of carrying sufficient energy to induce feedback in the host galaxy. The overall wind structure in this source, with heavy and light UFOs, represents a new example that ultra fast winds with a stratified structure are commonplace in Narrow Line Seyfert 1 galaxies, therefore providing a new candidate where the UFO may arise from a shock of the wind with the ISM.
Acknowledgment
The authors of this work thank the anonymous referee for his/her valuable comments. We thank the Instituto de Astronomía at Universidad Nacional Autónoma de México (IAUNAM). ALL, OSM and CV-C acknowledge support from DGAPA-PAPIIT IA103625. Lara-DI acknowledges support from Secretaria de Ciencia, Humanidades, Tecnología e Innovación (SECIHTI) through the Estancias Posdoctorales por México Convocatoria 2025 (Estancia Postdoctoral Iniciales) grant program. M.G. acknowledges financial support from CONAHCYT through the SNII Research Assistant programme, granted to support the research activities of A.L.L., a Level III National Researcher. MSL acknowledges support from PID2022-136598NB-C33 by MCIN/AEI/10.13039/501100011033.
References
- [1]Akaike, H. 1974, IEEE Transactions on Automatic Control, 19, 716
- [2]Bianchi, S., La Franca, F., Matt, G., Guainazzi, M., Jimenez Bailón, E., Longinotti, A. L., Nicastro, F., & Pentericci, L. 2008, MNRAS, 389, L52
- [3]Boller, T., Balestra, I., & Kollatschny, W. 2007, A&A, 465, 87
- [4]Boller, T., Brandt, W. N., & Fink, H. 1996, A&A, 305, 53
- [5]Boroson, T. A. & Green, R. F. 1992, ApJS, 80, 109
- [6]Bower, R. G., Benson, A. J., Malbon, R., Helly, J. C., Frenk, C. S., Baugh, C. M., Cole, S., & Lacey, C. G. 2006, MNRAS, 370, 645
- [7]Brandt, W. N., Mathur, S., & Elvis, M. 1997, MNRAS, 285, L25
- [8]Cash, W. 1979, ApJ, 228, 939
- [9]Chartas, G., Saez, C., Brandt, W. N., Giustini, M., & Garmire, G. P. 2009, ApJ, 706, 644
- [10]Di Matteo, T., Springel, V., & Hernquist, L. 2005, Nature, 433, 604
- [13]Feruglio, C., Fiore, F., Carniani, S., Piconcelli, E., Zappacosta, L., Bongiorno, A., Cicone, C., Maiolino, R., Marconi, A., Menci, N., Puccetti, S., & Veilleux, S. 2015, A&A, 583, A99
- [14]Fields, D. L., Mathur, S., Pogge, R. W., Nicastro, F., & Komossa, S. 2005a, ApJ, 620, 183
- [16]Gallo, L. C., Blue, D. M., Grupe, D., Komossa, S., & Wilkins, D. R. 2018, MNRAS, 478, 2557
- [17]Gofford, J., Reeves, J. N., Tombesi, F., Braito, V., Turner, T. J., Miller, L., & Cappi, M. 2013, MNRAS, 430, 60
- [18]Grupe, D. & Mathur, S. 2004, ApJ, 606, L41
- [19]Grupe, D., Thomas, H. C., & Beuermann, K. 2001, A&A, 367, 470
- [20]Gupta, A., Mathur, S., & Krongold, Y. 2015, ApJ, 798, 4
- [21]Gupta, A., Mathur, S., Krongold, Y., & Nicastro, F. 2013, ApJ, 772, 66
- [22]Harrison, C. M., Alexander, D. M., Mullaney, J. R., & Swinbank, A. M. 2014, MNRAS, 441, 3306
- [23]Harrison, C. M., Costa, T., Tadaki, C. N., Flütsch, A., Kakkad, D., Perna, M., & Vietri, G. 2018, Nature Astronomy, 2, 198
- [24]Harrison, F. A., Craig, W. W., Christensen, F. E., Hailey, C. J., Zhang, W. W., Boggs, S. E., Stern, D., Cook, W. R., Forster, K., Giommi, P., Grefenstette, B. W., Kim, Y., Kitaguchi, T., Koglin, J. E., Madsen, K. K., Mao, P. H., Miyasaka, H., Mori, K., Perri, M., Pivovaroff, M. J., Puccetti, S., Rana, V. R., Westergaard, N. J., Willis, J., Zoglauer, A., An, H., Bachetti, M., Barrière, N. M., Bellm, E. C., Bhalerao, V., Brejnholt, N. F., Fuerst, F., Liebe, C. C., Markwardt, C. B., Nynka, M., Vogel, J. K., Walton, D. J., Wik, D. R., Alexander, D. M., Cominsky, L. R., Hornschemeier, A. E., Hornstrup, A., Kaspi, V. M., Madejski, G. M., Matt, G., Molendi, S., Smith, D. M., Tomsick, J. A., Ajello, M., Ballantyne, D. R., Baloković, M., Barret, D., Bauer, F. E., Blandford, R. D., Niel Brandt, W., Brenneman, L. W., Chiang, J., Chakrabarty, D., Chenevez, J., Comastri, A., Dufour, F., Elvis, M., Fabian, A. C., Farrah, D., Fryer, C. L., Gotthelf, E. V., Grindlay, J. E., Helfand, D. J., Krivonos, R., Meier, D. L., Miller, J. M., Natalucci, L., Ogle, P., Ofek, E. O., Ptak, A., Reynolds, S. P., Rigby, J. R., Tagliaferri, G., Thorsett, S. E., Treister, E., & Urry, C. M. 2013, ApJ, 770, 103
- [25]Hebbar, P. R., Heinke, C. O., Sivakoff, G. R., & Shaw, A. W. 2019, MNRAS, 485, 5604
- [26]Hopkins, P. F. & Elvis, M. 2010, MNRAS, 401, 7
- [27]Juránová, A., Costantini, E., Di Gesu, L., Ebrero, J., Kaastra, J., Korista, K., Kriss, G. A., Mehdipour, M., Piconcelli, E., & Rogantini, D. 2024, A&A, 681, A101
- [28]Kalberla, P. M. W., Burton, W. B., Hartmann, D., & Arnal, E. M., Bajaja, E., Morras, R., & Pöppel, W. G. L. 2005, A&A, 440, 775
- [29]King, A. R. 2010, MNRAS, 402, 1516
- [30]Kollatschny, W. 2003, A&A, 407, 461
- [31]Kollatschny, W., Zetzl, M., & Dietrich, M. 2006, A&A, 454, 459
- [32]Kormendy, J. & Ho, L. C. 2013, ARA&A, 51, 511
- [33]Kosec, P., Buisson, D. J. K., Parker, M. L., Pinto, C., Fabian, A. C., & Walton, D. J. 2018, MNRAS, 481, 947
- [34]Krongold, Y., Longinotti, A. L., Santos-Lleo, M., Mathur, S., Peterson, B. M., Nicastro, F., Gupta, A., Rodríguez-Pascual, P., & Elías-Chávez, M. 2021, ApJ, 917, 39
- [35]Longinotti, A. L., Bianchi, S., Santos-Lleo, M., Rodríguez-Pascual, P., Guainazzi, M., Cardaci, M., & Pollock, A. M. T. 2007, A&A, 470, 73 Longinotti, A. L., Krongold, Y., Guainazzi, M., Giroletti, M., Panessa, F., Costantini, E., Santos-Lleo, M., & Rodriguez-Pascual, P. 2015, ApJ, 813, L39
- [36]Longinotti, A. L., Salome, Q., Feruglio, C., Krongold, Y., Garcia-Burillo, S., Giroletti, M., Panessa, F., Stanghellini, C., Vega, O., Patino-Alvarez, V. M., Chavushyan, V., Elias-Chavez, M., & Robleto-Orus, A. 2023, MNRAS, 521, 2134
- [37]Longinotti, A. L., Vega, O., Krongold, Y., Aretxaga, I., Yun, M., Chavushyan, V., Feruglio, C., Gomez-Ruiz, A., Montaña, A., Leon-Tavares, J., Olguin-Iglesias, A., Giroletti, M., Guainazzi, M., Kotilainen, J., Panessa, F., Zapata, L. A., Cruz-Gonzalez, I., Patino-Alvarez, V. M., Rosa-Gonzalez, D., Carramiñana, A., Carrasco, L., Costantini, E., Dultzlin, D., Guichard, J., Puerari, I., & Santos-Lleo, M. 2018, ApJ, 867, L11
- [38]Luminari, A., Nicastro, F., Elvis, M., Piconcelli, E., Tombesi, F., Zappacosta, L., & Fiore, F. 2021, A&A, 646, A111
- [39]Marasco, A., Cresci, G., Nardini, E., Mannucci, F., Marconi, A., Tozzi, P., Tozzi, G., Amiri, A., Venturi, G., Piconcelli, E., Lanzuisi, G., Tombesi, F., Mingozzi, M., Perna, M., Carniani, S., Brusa, M., & di Serego Alighieri, S. 2020, A&A, 644, A15
- [40]Mathur, S. 2000, MNRAS, 314, L17
- [41]Mathur, S., Fields, D., Peterson, B. M., & Grupe, D. 2012, ApJ, 754, 146
- [42]Mathur, S. & Grupe, D. 2005a, A&A, 432, 463
- [46]Mizumoto, M., Reeves, J. N., Braito, V., Behar, E., Done, C., Hagino, K., Kraemer, S. B., Matzeu, G. A., Noda, H., Nomura, M., Ogawa, S., Ohsuga, K., Tanimoto, A., Turner, T. J., Ueda, Y., Yamada, S., Ganguly, S., & Somenzi, P. 2026, ApJ, 997, 219
- [47]Osterbrock, D. E. & Pogge, R. W. 1985, ApJ, 297, 166
- [48]Panessa, F., Perez-Torres, M., Hernandez-Garcia, L., Casella, P., Giroletti, M., Orienti, M., Baldi, R. D., Bassani, L., Fiocchi, M. T., La Franca, F., Malizia, A., McHardy, I., Nicastro, F., Piro, L., Vincentelli, F., Williams, D. R. A., & Ubertini, P. 2022, MNRAS, 510, 718
- [49]Parker, M. L., Pinto, C., Fabian, A. C., Lohfink, A., Buisson, D. J. K., Alston, W. N., Kara, E., Cackett, E. M., Chiang, C.-Y., Dauser, T., De Marco, B., Gallo, L. C., Garcia, J., Harrison, F. A., King, A. L., Middleton, M. J., Miller, J. M., Miniutti, G., Reynolds, C. S., Uttley, P., Vasudevan, R., Walton, D. J., Wilkins, D. R., & Zoghbi, A. 2017, Nature, 543, 83
- [50]Piconcelli, E., Jimenez-Bailon, E., Guainazzi, M., Schartel, N., Rodriguez-Pascual, P. M., & Santos-Lleo, M. 2004, MNRAS, 351, 161
- [51]Pinto, C., Alston, W., Parker, M. L., Fabian, A. C., Gallo, L. C., Buisson, D. J. K., Walton, D. J., Kara, E., Jiang, J., Lohfink, A., & Reynolds, C. S. 2018, MNRAS, 476, 1021
- [52]Pinto, C., Alston, W., Soria, R., Middleton, M. J., Walton, D. J., Sutton, A. D., Fabian, A. C., Earnshaw, H., Urquhart, R., Kara, E., & Roberts, T. P. 2017, MNRAS, 468, 2865
- [53]Porquet, D., Reeves, J. N., Grosso, N., Braito, V., & Lobban, A. 2021, A&A, 654, A89
- [54]Pounds, K. A. 2014, MNRAS, 437, 3221
- [55]Pounds, K. A. & Vaughan, S. 2011, MNRAS, 413, 1251
- [56]Rees, M. J. 1984, ARA&A, 22, 471
- [57]Reeves, J. N. & Braito, V. 2019, ApJ, 884, 80
- [58]Reeves, J. N., Braito, V., Nardini, E., Hamann, F., Chartas, G., Lobban, A. P., O’Brien, P. T., & Turner, T. J. 2018, ApJ, 867, 38
- [59]Reeves, J. N., Porquet, D., Braito, V., Grosso, N., & Lobban, A. 2021, A&A, 649, L3
- [60]Sanfrutos, M., Longinotti, A. L., Krongold, Y., Guainazzi, M., & Panessa, F. 2018, ApJ, 868, 111
- [61]Tombesi, F., Melendez, M., Veilleux, S., Reeves, J. N., González-Alfonso, E., & Reynolds, C. S. 2015, Nature, 519, 436
- [62]Tombesi, F., Sambruna, R. M., Reeves, J. N., Braito, V., Ballo, L., Gofford, J., Cappi, M., & Mushotzky, R. F. 2010, ApJ, 719, 700
- [63]Watson, L. C., Mathur, S., & Grupe, D. 2007, AJ, 133, 2435
- [64]Wilms, J., Allen, A., & McCray, R. 2000, ApJ, 542, 914
- [65]Xiang, X., Miller, J. M., Behar, E., Boissay-Malaquin, R., Brenneman, L., Buhariwalla, M., Byun, D., Done, C., Gallo, L., Geronemato, D., Hagen, S., Kaastra, J., Paltani, S., Porter, F. S., Mushotzky, R., Noda, H., Mehdipour, M., Minezaki, T., Tashiro, M., & Zoghbi, A. 2025, ApJ, 988, L54
- [67]Zaidouni, F., Kara, E., Kosec, P., Behar, E., Mushotzky, R., Koss, M., Juráňová, A., Kammoun, E., Brenneman, L. W., Chakraborty, J., Ebisawa, K., Eckart, M. E., Fabian, A. C., Fukazawa, Y., Garcia, J. A., Gu, L., Masterson, M., Ogawa, S., Okajima, T., Paltani, S., Rogantini, D., Terashima, Y., Williams, B. J., & Yamada, S. 2026, ApJ, 997, L26