WorksPhysics and Astronomy
The Black Hole Mass of the Sunflower Galaxy NGC 5055: Direct Measurement via Stellar Kinematics with Keck/OSIRIS Laser Guide Star Adaptive Optics
Abstract
We present the first stellar-dynamical supermassive black hole mass measurement in the nearby spiral galaxy NGC 5055 (M63) using near-infrared integral-field spectroscopy with Keck/OSIRIS assisted by the laser guide star adaptive optics system. The observations deliver two-dimensional stellar kinematics extracted from the CO bandhead absorption features via pPXF at an angular resolution of . The stellar velocity dispersion rises steeply from at to a central peak of , kinematically tracing the gravitational influence of a central mass concentration. Modeling these kinematics with an ensemble of four JAM models, spanning cylindrical and spherical velocity-ellipsoid alignments and two independent MGE decompositions of archival HST imaging, yields a black hole mass of . The stellar orbits are consistent with near-isotropy (), typical of rotating late-type galaxies. NGC 5055 emerges as a remarkable local analogue to the Milky Way in terms of its black hole, nuclear star cluster, and total stellar mass. The surface brightness profile of NGC 5055 shows no evidence for a classical bulge; if confirmed bulgeless, NGC 5055 would join the rare class of spirals with a direct dynamical determination in the absence of a bulge, a regime currently populated by only a handful of galaxies including NGC 2748 and NGC 4395. This measurement provides a new anchor for black hole scaling relations in late-type galaxies at the low-mass end of the black hole mass spectrum.
Introduction
While morphology-dependent coevolutionary scaling relations between supermassive black holes (SMBHs) and their host galaxies have emerged in recent years, this fundamental paradigm would benefit from more data at the low-mass end. In the regime of late-type and dwarf galaxies with , the precise mechanisms driving BH-galaxy co-evolution remain a subject of active and vigorous research. The tight correlations discovered between black hole (BH) mass and host-galaxy properties, such as the stellar velocity dispersion within the inner region () or the effective half-light radius (), and the bulge luminosity (), stellar mass , dynamical mass, and radial concentration of stars (e.g., L. Ferrarese & D. Merritt [40]; K. Gebhardt et al. [42]; A. W. Graham et al. [53]; R. J. McLure & J. S. Dunlop [83]; A. Marconi & L. K. Hunt [79]; A. W. Graham [45]), suggest one, or multiple, fundamental co-evolutionary links between SMBHs and their hosts over cosmic time.
Active galactic nuclei (AGN) feedback models propose that SMBHs regulate host-galaxy star formation via momentum- or energy-driven winds, providing a physical basis for part of these scaling relations (M. G. Haehnelt et al. [57]; J. Silk & M. J. Rees [118]; A. King [66]; T. Di Matteo et al. [34]). However, the precise mechanisms driving this co-evolution remain a subject of active debate and may involve both energy and momentum conserving winds at different stages (J. Silk et al. [117]). At high masses, A. W. Graham [48] explains how (cold gas)-poor mergers drive a dramatic departure, as dictated by the Virial theorem, from these relations predicted by AGN feedback. At low masses, there is reason to believe that galaxies again deviate from the AGN-dictated scaling relations. Indeed, given that dwarf early-type galaxies (dETGs) follow an relation (e.g., R. L. Davies et al. [29]; E. Toloba et al. [127]), then if they follow a near-linear - relation (e.g., J. Magorrian et al. [76]), they must follow an - relation, placing them above-and-left of the extrapolated AGN-expectation for - relation at low-masses. At the same time, some spiral galaxies have been observed to reside to the lower-right of this relation (A. W. Graham [44]; J. Hu [62]). Expanding the sample of direct, dynamical measurements at low masses is therefore essential for discriminating between competing theoretical models and for calibrating secondary mass indicators (see L. Ferrarese & H. Ford [39]; J. Kormendy & L. C. Ho [67]; A. W. Graham [46], for reviews).
Late-type galaxies (LTGs) are particularly valuable in this context. They span a wide range of morphologies, structural complexities, and star-formation histories, and they preferentially occupy the low-mass end of the distribution, precisely the regime where competing co-evolution models make divergent predictions. Moreover, LTGs frequently host composite central structures shaped by both merger-driven and secular processes, making them sensitive probes of how –host relations depend on assembly history. Yet among the direct measurements published to date (e.g., A. W. Graham [44]; J. Kormendy & L. C. Ho [67]; R. P. Saglia et al. [112]; R. C. E. van den Bosch 2016; N. Sahu et al. [113]), only several dozen reside in LTGs, introducing a significant morphological bias into the current census.
Obtaining dynamical in LTGs is observationally challenging for two primary reasons. First, the predicted from scaling relations is relatively small (– for ), such that the gravitational sphere of influence of the BH within which its gravitational potential dominates the stellar dynamics, defined as , where the enclosed stellar mass equals approximately .1 Second, the prominent disk and dust features in LTGs complicate optical surface-brightness photometry and can bias dynamical mass models. These obstacles are increasingly overcome through (i) near-infrared (NIR) integral-field spectroscopy (IFS) with laser-guide-star adaptive optics (LGS-AO), which provides resolution while penetrating dust extinction; and (ii) Jeans Anisotropic Modeling (JAM)2; M. Cappellari [15, 17] paired with Multi-Gaussian Expansion (MGE)3; E. Emsellem et al. [38]; M. Cappellari [14] photometric modeling. JAM efficiently models two-dimensional (2D) stellar kinematic maps using axisymmetric Jeans equations that account for a central dark object, orbital anisotropy, and the bulk of the galaxy’s mass distribution. Such methods have yielded robust measurements in ETGs (e.g., D. D. Nguyen et al. [94, 95, 96, 92, 102]; S. Thater et al. [123, 124]) and, more recently, in a growing number of LTGs (e.g., M. Cappellari et al. [22]; D. A. Drehmer et al. [36]; D. D. Nguyen et al. [100, 103]; T. Q. T. Le et al. 2026).
The Target NGC 5055
NGC 5055 (M63, aka the “Sunflower Galaxy”) is a nearby Sbc spiral galaxy in the small M51 Group of galaxies. NGC 5055 is mag fainter than M51 (aka the “Whirlpool Galaxy”), with the next most dominant galaxy (NGC 5023, if ignoring M51B) some 3.6 mag fainter. With two dominant spiral galaxies, it therefore has some similarity to the Local Group, which is dominated by M31+M32 and the Milky Way.
NGC 5055 has a flocculent appearance at optical wavelengths (D. M. Elmegreen & B. G. Elmegreen [37]) but displays a weak grand-design two-arm spiral () in the -band (M. D. Thornley & L. G. Mundy [126]). Signs of minor disturbances are evident in optical images from, for example, the Virtal Telescope Project4 (G. Masi [82]) and extensive tidal tails and asymmetries in the galaxy emerge in deep images (D. Martínez-Delgado et al. [81]; G. Morales et al. [88]). NGC 5055 has a precisely determined Tip-of-the-Red-Giant-Branch (TRGB) distance of (K. B. W. McQuinn et al. [84]), corresponding to a physical scale of 43 pc arcsec under a flat CDM cosmology with , , and . At this distance, the logarithm of the black hole mass is predicted to be dex (), when using the – relation for LTGs ([113 their equation 2]) and ((L. C. Ho et al. 2009)). This yields R_{\mathrm{SOI}} \simeq0\farcs069 (= 2.96 pc), a factor of 1.5 above the resolution limit of our Keck/OSIRIS LGS-AO data, which has a full width at half maximum of the point spread function of \mathrm{FWHM}_{\mathrm{PSF}}^{\mathrm{OSIRIS}} = 0\farcs045 (Table 1).
| (Light fraction) | (arcsec) | FWHM (arcsec) | |
| 1 | 0.805 | 0.017 | |
| 2 | 0.187 | 0.040 | 0.045 |
| 3 | 0.008 | 0.150 |
Table 1. Keck/OSIRIS PSF MGE models (HIP 61471). Notes. (1) Gaussian component index; (2) fractional contribution to the total light; (3) Gaussian dispersion along the major axis; (4) intensity-weighted mean FWHM of the PSF.
Previous attempts to constrain the central mass in NGC 5055 lacked the spatial resolution to isolate the BH’s gravitational influence from the surrounding stellar potential. I. Perez et al. [109] derived a total dynamical enclosed mass of within the central ( pc) from ground-based long-slit spectroscopy, also revealing a linear rotation curve and a outflow in this region. From CFHT Fabry-Perot H observations, S. Blais-Ouellette et al. [12] identified anomalous kinematics consistent with a compact rotating disk of radius \sim8\farcs5 ( pc) and inferred a central dark mass of via a simple Keplerian estimate. While these two estimates are not contradictory: the former is the total enclosed mass dominated by stars and gas, the latter a point-like dark object, neither resolves the BH SOI, which is expected to extend only to pc (§ 6). No direct stellar-dynamical measurement has since been reported for NGC 5055.
The nuclear region is otherwise well characterized. The nucleus is classified as a low-ionization nuclear emission-line region galaxy (LINER; [59, 60]) and exhibits signatures of low-luminosity AGN activity, including compact X-ray emission and UV stellar-wind features ([78, 77]). IFS with the 6 m SAO/RAS telescope by V. L. Afanasiev & O. K. Sil’chenko [1] identified a kinematically distinct circumnuclear stellar disk and a lopsided, Mg-enhanced substructure displaced pc () south-west of the photometric center, with mean stellar ages of 3–4 Gyr, significantly younger than the surrounding populations (8–10 Gyr), suggesting episodic star formation driven by secular gas inflow. On larger scales, H I interferometry reveals a symmetric, warped gaseous disk extending to kpc ([8]), while the Spitzer Survey of Stellar Structure in Galaxies (S4G) photometry provides a well-constrained stellar mass-to-light ratio for dynamical modeling (; [65]). NGC 5055 thus represents a nearby, multiwavelength-rich target lacking a stellar-dynamical determination.
Crucially, with a stellar mass of ([30]), NGC 5055 provides an opportunity to study a local, late-type structural analogue to the Milky Way ( dex), allowing us to better place our own galaxy’s evolutionary history into a broader cosmological context.
This Work
We present the first stellar-dynamical measurement of in NGC 5055 using NIR IFS observations obtained with OSIRIS behind the LGS-AO system on the 10-m Keck I telescope. With a PSF FWHM of 0\farcs045, our data achieve a physical resolution of pc, providing the highest-resolution view of this galaxy’s central region to date. Two-dimensional line-of-sight (LOS) velocity distribution (LOSVD) of the stellar kinematics is extracted from the CO absorption bandheads () using the Penalized PiXel-Fitting (pPXF5) method ([21, 16]). These kinematics are modeled using JAM ([15, 17]) in conjunction with MGE decompositions of archival Hubble Space Telescope (HST) and ground-based -band imaging that characterize the surface brightness profile from pc to kpc scales ([41, 108]).
This paper is organized as follows. § 2 describes the Keck/OSIRIS observations, data reduction, and PSF characterization. § 3 and § 4 present the stellar kinematic extraction and photometric modeling of NGC 5055, respectively. § 5 describes the JAM dynamical modeling. Results and discussion are presented in § 6, and we summarize our conclusions in § 7.
OSIRIS Observations and Data Reduction
IFS Observations and Data Reduction
We observed the nucleus of NGC 5055 with the OH Suppressing InfraRed Imaging Spectrograph (OSIRIS; [70]) with the Keck I LGS-AO system ([137] (M. A. van Dam et al. 2006)). OSIRIS is a lenslet array IFU spectrograph with a 2048 2048 Hawaii-2 detector and spectral resolution in the band of . The observations were conducted on 15 Jan. 2025, and consisted of two 900-second object frames, with an intervening 900-second sky frame (i.e., 30 min on object), in the standard object-sky-object sequence, at an airmass of 1.08. The plate scale was 0\farcs035, which at the distance of NGC 5055 (8.87 Mpc) equates to 1.5 pc pixel. We observed with the filter (2.292–2.408 m), which is ideally suited for CO bandhead observations.
Each object frame was reduced using the OSIRIS data reduction pipeline (DRP) version 6.0, using the sky frames for background subtraction. The DRP uses “Rectification Matrices” which incorporate all the flat fielding, bias and dark frames in one, along with mapping each pixel to a lenslet/spectral element. These matrices are retrieved from the Keck repository; for our data, these were dated Feb. 2025.
The resulting data cubes have fields of view of 1\farcs785 \times2\farcs31 ( pixels).
The data cubes were centered and combined, then trimmed for display purposes, resulting in images with about 2100 spectra. The resulting data cube values were cleaned of spikes and geometric artifacts; these are mainly a product of the DRP. Telluric correction was performed with observation of the A0V star, HIP 61471. The stellar spectrum is interpolated over the hydrogen absorption features and a representative blackbody is assumed. This is divided through the data cube spectra to remove the telluric absorption features. Flux calibration is performed by integrating the stellar spectrum over a clean region (away from telluric and stellar features), then matched to a reference Vega spectrum.
OSIRIS Point Spread Functions
Since the empirical OSIRIS PSF depends on the AO correction quality and atmospheric conditions at the time of observation, we derived it directly from a standard star observed contemporaneously with the NGC 5055 science frames. We observed HIP 61471 with the AO loops closed immediately following the science exposures in a short exposure time to avoid saturation, using the same plate scale ( mas pixel) and filter (). The PSF image was constructed from the normalized, median-combined data cube collapsed over –m, a wavelength range within the bandpass and matching the spectral region used for stellar kinematic extraction. This empirical PSF thus captures the instantaneous AO performance during the NGC 5055 observations.
The PSF image was parameterized as a Multi-Gaussian Expansion (MGE; E. Emsellem et al. [38]) using the mge_fit_sectors routine from the MgeFit6 package (M. Cappellari [14]). The resulting MGE model is listed in Table 1 and is convolved with all JAM model predictions to account for the OSIRIS PSF at m (§ 5). The azimuthally averaged radial profile of HIP 61471 and this corresponding MGE fit are shown in Figure 1, with the PSF FWHM and SOI indicated for reference.
Figure 1. Keck/OSIRIS PSF constructed from the star HIP 61471, which is extracted along the major (left) and minor (right) axes. Black open circles show the data and the red line is its best-fit 2D MGE model (Table 1). At , the MGE fit reproduces the PSF profile to within 8% uncertainty.
Stellar Kinematics Measurements
Defining the Galaxy Center
Due to the uncertainties in LOSVD extractions (§ 3.3), we adopt the photometric center as the coordinate origin, assuming the central BH coincides with the surface-brightness peak in the OSIRIS white-light image. To determine the weighted barycenter of the nucleus, we employed a two-step refinement: first using the 200 brightest pixels within a 0\farcs35 \times0\farcs7 aperture of the OSIRIS white-light image, then the 60 brightest pixels of its 0\farcs21 \times0\farcs35 region. The estimates agree within a quarter of a pixel (\Delta X < 0\farcs018, \Delta Y < 0\farcs017), a difference negligible at the spatial resolution of the Keck/OSIRIS IFU data. Consequently, we adopt the barycenter from the larger aperture (R.A., Decl. = 13^{\mathrm h}15^{\mathrm m}49\farcs155, +42^\circ01'46\farcs86; consistent with the HST/WFPC2 F547M photometric center from R. Pechetti et al. [108] determination for its superior statistical robustness and used it as the origin in all subsequent models and figures.
pPXF Setup & The Optimal Stellar Template
Stellar kinematics for NGC 5055 were extracted from the Keck/OSIRIS -band data cube using the pPXF method (M. Cappellari [18]), primarily targeting the CO bandhead absorption features (e.g., M. Cappellari et al. [22]; D. D. Nguyen et al. [93]; H. N. Ngo et al. 2025). Following the results of D. D. Nguyen et al. [103], who found that LOSVDs derived from the empirical X-shooter Spectral Library (XSL7; K. Verro et al. [133]) agree within with those from the higher-resolution PHOENIX synthetic library (T.-O. Husser et al. [64]), we adopt XSL as our primary template source. The XSL provides broad spectral coverage (– Å) at , encompassing a diverse range of O–M and evolved stars.
To account for mixed stellar populations in the nucleus of NGC 5055 and mitigate kinematic biases from template mismatch (A. Marconi et al. [80]; J. D. Silge & K. Gebhardt [116]), we constructed an optimal XSL template from a large spatial region’s integrated spectrum, namely the global spectrum. Additive Legendre polynomials were excluded from all pPXF fits (degree ), as they can modify the equivalent widths of the CO bandhead absorption features in the -band and systematically bias the recovered LOSVDs. Multiplicative polynomials (mdegree) of orders – were instead explored to account for residual continuum shape, sky subtraction, and calibration systematics; the recovered kinematics converge and become insensitive to polynomial order at mdegree over the full fitting window (–m). We therefore adopt mdegree as our fiducial configuration.
The procedure is as follows. First, we constructed a global spectrum by co-adding all spaxels within the red aperture shown in the left panel of Figure 2 from the OSIRIS data cube, weighted by the continuum flux of each spaxel. The resulting spectrum reaches a signal-to-noise ratio of S/N per spectral pixel and was logarithmically rebinned to a constant velocity scale of 32 km s per pixel following Eq. 8 of M. Cappellari [16]. We broadened the XSL templates to match the instrumental resolution by convolving them with a Gaussian kernel with dispersion computed from Eq. 5 of M. Cappellari [16].
Figure 2. Left: Logarithmic map of the integrated stellar continuum from the Keck/OSIRIS data cube, collapsed along the spectral axis. Right: Optimal stellar template fit to the integrated spectrum of NGC 5055 by co-adding all spaxels within the red aperture shown in the left panel. The observed spectrum (black) is overlaid with the best-fit pPXF broadened template (red); residuals are shown as green dots. White contours trace the stellar continuum intensity in steps of 1 mag arcsec.
Second, we derived a fixed optimal XSL template by fitting the broadened templates to the global spectrum with pPXF, adopting a Gaussian LOSVD (i.e., moments = 2). This yields the stellar kinematics, namely the rotational velocity (relative to the systemic velocity ) and the velocity dispersion , which are the primary quantities of interest.
The optimal template consists of 15 giant-star spectra that reproduce the observed features over 2.29–2.40 m, including the CO bandhead absorption (Figure 2, right) of the OSIRIS spectra. The fit is dominated by HD 161817 (XSL: X0639), contributing of the flux, with additional contributions from a giant K-type star (HD 38529B; ) and an M-type star (BMB 106; ). The best-fit integrated stellar kinematics yield a heliocentric velocity km s and velocity dispersion km s, where uncertainties represent the scatter from 200 Monte Carlo realizations [21]. Both values are in good agreement with independent measurements: is consistent with the H I 21 cm systemic velocity of C. M. Springob et al. [122], and agrees with the CaT and Mg Ib absorption-line dispersion measured with the Double Spectrograph at Palomar Observatory [7].
Two-Dimensional Stellar Kinematic Maps
Two-dimensional stellar LOSVDs across the nucleus of NGC 5055 were extracted as follows. We first applied the adaptive PowerBin8 algorithm [19] to spatially bin spaxels to a target S/N of 162 per spectral pixel, yielding bins. Spaxels beyond were grouped adaptively, while those interior to were
left unbinned to preserve the spatial resolution required for kinematic measurements within the SOI; individual unbinned spaxels there have S/N ranging from 162 to 165 per spectral pixel, increasing toward the nucleus. Each binned spectrum was resampled onto a logarithmic wavelength grid and fitted with pPXF (M. Cappellari [16]) using the fixed optimal XSL template (§ 3.2) broadened to the instrumental resolution, with second-degree multiplicative Legendre polynomials to correct for residual spectral shape differences, yielding the stellar LOS velocity and dispersion .
Figure 3 presents the logarithmic integrated intensity map (panel A), with a marked spaxel corresponding to the spectrum and pPXF fit shown in panel B. The remaining panels display the resulting maps of rotational velocity (panel C), velocity dispersion (panel D), and root-mean-square velocity (panel E). The nucleus exhibits a rotation with km s at the edge of the OSIRIS FOV (after subtracting the systemic heliocentric velocity of km s). This derived from the Keck/OSIRIS data cube is identical to the value reported in the NASA/IPAC Extragalactic Database (NED)9.
Figure 3. Stellar kinematics of NGC 5055 from Keck/OSIRIS. Panel A: Logarithmic integrated intensity map of the data cube. Panel B: Representative pPXF fit for a central-offset PowerBin bin (location marked in Panel A). The observed spectrum (black) shows CO bandhead absorptions near 2.3 m; the best-fit XSL template is overplotted in red. Residuals data–model are shown in green; vertical gray lines delimit the fitting window applied uniformly across all bins. Panels C–E: 2D maps of the stellar line-of-sight velocity (), velocity dispersion (), and root-mean-square velocity (). White contours trace the stellar continuum intensity in steps of 1 mag arcsec.
We measure a global kinematic position angle of . Both and were derived from the velocity map using the PAFIT10 package (D. Krajnović et al. [68]). The kinematic is consistent within the uncertainties with the photometric one reported in the HyperLeda database11 (B. M. H. R. Wevers et al. [136]), as expected for a nearly axisymmetric disk galaxy. The profile, which dominates over , rises from km s at , to a central peak of km s within . This central increase in both and provides strong evidence for the presence of a central SMBH. We presented these stellar-kinematic measurements in Table 2.
| R.A. (arcsec) | Decl. (arcsec) | (km s) | (km s) | (km s) | (km s) |
| 0.001 | 3.47 | 134.67 | 3.23 | ||
| 0.018 | 14.73 | 3.04 | 126.98 | 2.97 | |
| 0.018 | 0.052 | 3.00 | 126.07 | 3.01 | |
| 0.96 | 2.96 | 126.06 | 2.93 | ||
| 0.059 | 6.14 | 3.13 | 121.43 | 3.21 | |
| 0.069 | 3.23 | 126.02 | 2.99 | ||
| 23.98 | 2.53 | 121.75 | 2.73 | ||
| ... | ... | ... | ... | ... | ... |
Table 2. Keck/OSIRIS Kinematic Data of the NGC 5055 Nucleus. (This table is available in its entirety in machine-readable form)
Galaxy Photometric Model
We retrieved the D. B. Fisher & N. Drory [41] composite -band surface brightness profile of NGC 5055, i.e., the central HST/WFPC2 PC F606W profile mated to the wide-field Sloan Digital Sky Survey (SDSS) profile (D. G. York et al. [138]). We performed an independent decomposition using a Python 3 version of the PROFILER code (B. C. Ciambur [27])12 to fit a three-component parametric model: two Sérsic functions plus an anti-truncated (broken) exponential disk, chosen to provide a good empirical match to the observed profile. The best-fit parameters, shown in Figure 4, are: an inner Sérsic component with mag arcsec, , and ; a middle Sérsic component with mag arcsec, , and (held fixed); and an anti-truncated exponential disk with central surface brightness mag arcsec, break radius r_b = 258\farcs9, and inner and outer scale lengths h_1 = 83\farcs5 and h_2 = 123\farcs8, respectively. Since all fitted components have effective radii well exceeding the WFPC2/PC PSF FWHM (\sim0\farcs1), PSF-broadening has a negligible effect on these reported parameters. The overall fit quality is characterized by a root-mean-square residual of mag arcsec, confirming that the adopted parametric model provides an adequate empirical description of the observed light profile. This decomposition improves upon D. B. Fisher & N. Drory [41] in two respects: it explicitly includes the compact inner component they excluded from their fit, and it adopts an anti-truncated disk model to capture the observed upturn in the profile beyond .

Figure 4. Decomposition of the surface brightness profile of NGC 5055 using the composite -band major-axis profile from D. B. Fisher & N. Drory [41], that combined HST/WFPC2 PC F606W imaging in the central region with wide-field ground-based photometry at large radii. Top: The observed profile (filled circles) is plotted against and overlaid with a three-component model (solid black line), comprising an inner Sérsic component (short-dashed red curve), a middle Sérsic component (dot-dashed red curve), and an anti-truncated exponential disk (solid blue curve) that dominates at large radii. Bottom: Residuals () in mag arcsec.

Figure 5. Stellar kinematic profiles of NGC 5055 extracted from the Keck/OSIRIS data cube (filled circles), overlaid with the best-fit JAM (left) and JAM (right) model predictions (solid curves). Red and blue points correspond to the major- and minor-axis bins, respectively, with the major-axis profile offset by +30 km s for clarity. Best-fit parameters and their 1 uncertainties are listed in the legend of each panel.
We stress that these Sérsic components serve purely as a flexible empirical parameterization of the light profile to enable MGE construction; they do not carry a physical interpretation as distinct structural components, for which we refer to § 6.4. The morphological classification of NGC 5055’s central region remains debated in the literature; D. B. Fisher & N. Drory [41] consider it to have a pseudobulge with (dominating from to ), while V. L. Afanasiev & O. K. Sil’chenko [1] favor a classical bulge interpretation (within ) based on the near- inner profile. Although, both studies acknowledge complicating factors including a circumnuclear stellar disk, dust lanes, and nuclear spiral structure. Given that V. L. Afanasiev & O. K. Sil’chenko [1] note that the excess light over the inner few arcseconds stems from, among other things, a circumnuclear cumnuclear stellar disk (confirmed by Figure 4), we do not regard this excess to be a traditional bulge.
We adopt our three-component parametric fit to the D. B. Fisher & N. Drory [41] composite -band light profile as our fiducial stellar mass model, with the HST/WFPC2 F547M two-component fit of R. Pechetti et al. [108] serving as an independent check to bracket the systematic uncertainty from the choice of photometric data. In both cases, the best-fit parameters were converted into MGE models using the mge_fit_1d routine (M. Cappellari [14]), applied to the circularized 1D major-axis profile on a logarithmic radial grid, with the observed axial ratio of each Gaussian component assigned from the ellipticity measurements of R. Pechetti et al. [108]. We fit the MGE to the parametric model rather than the data directly, as the composite profile (HST + ground-based) becomes increasingly noisy at large radii and a smooth analytical representation avoids spurious Gaussian components in the MGE. Both MGE models comprise 17 Gaussian components and are listed in Table 3.
| (L pc) | (arcsec) | (L pc) | (arcsec) | |||
| (1) | (2) | (3) | (4) | (2) | (3) | (4) |
| R. Pechetti et al. [108] | Our refit | D. B. Fisher & N. Drory (2008) | ||||
| 1 | 636972 | 0.002 | 0.87 | 275042 | 0.050 | 0.87 |
| 2 | 575440 | 0.007 | 0.87 | 103922 | 0.115 | 0.87 |
| 3 | 407380 | 0.021 | 0.87 | 43763 | 0.225 | 0.61 |
| 4 | 224400 | 0.054 | 0.87 | 10564 | 0.345 | 0.61 |
| 5 | 65700 | 0.072 | 0.87 | 15921 | 0.490 | 0.61 |
| 6 | 92700 | 0.127 | 0.87 | 8129 | 0.880 | 0.61 |
| 7 | 39100 | 0.138 | 0.87 | 3182 | 1.522 | 0.61 |
| 8 | 9430 | 0.196 | 0.87 | 1551 | 2.498 | 0.61 |
| 9 | 28600 | 0.273 | 0.61 | 760 | 4.522 | 0.61 |
| 10 | 23500 | 0.276 | 0.61 | 409 | 9.920 | 0.61 |
| 11 | 15300 | 0.488 | 0.61 | 513 | 27.596 | 0.61 |
| 12 | 5830 | 0.621 | 0.61 | 23 | 32.443 | 0.61 |
| 13 | 4120 | 0.821 | 0.61 | 102 | 69.874 | 0.61 |
| 14 | 4290 | 0.957 | 0.61 | 7 | 82.874 | 0.61 |
| 15 | 3320 | 1.713 | 0.61 | 69 | 111.568 | 0.61 |
| 16 | 964 | 3.844 | 0.61 | 28 | 169.309 | 0.61 |
| 17 | 1130 | 6.296 | 0.61 | 3 | 247.063 | 0.61 |
Table 3. HST/WFPC2 PC F547M and F606W (plus -band ground-based photometry) MGE models. Notes: (1) Gaussian component index; (2) Luminosity surface density; (3) Gaussian dispersion along the major axis; (4) Observed axial ratio.
We note that the MGE parameterization of a surface brightness profile is mathematically non-unique: multiple combinations of Gaussian components can equally well reproduce the observed light distribution to within the measurement uncertainties. The two MGE models presented in Table 3 differ in their individual Gaussian parameters, reflecting this degeneracy rather than any physical distinction between the underlying stellar mass distributions. What matters for the dynamical modeling is that both models accurately reproduce the observed surface brightness profile; as demonstrated by the consistent values recovered across all four JAM models (Table 4), the dynamical results are robust against the specific choice of MGE decomposition.
| Parameter (Unit) (MGE) | Search range | Best-fit value JAM | Error (16–84)% | Error (0.14–99.86)% | Best-fit value JAM | Error (16–84)% | Error (0.14–99.86)% |
| Refit D. B. Fisher & N. Drory [41] | F606W | JAM | JAM | ||||
| (0 → 8) | 6.57 | −0.33, +0.18 | −1.52, +0.37 | 6.65 | −0.30, +0.14 | −1.33, +0.30 | |
| () | (0.1 → 3) | 1.75 | −0.10, +0.11 | −0.22, +0.31 | 1.65 | −0.20, +0.24 | −0.45, +0.70 |
| (0.05 → 0.99) | 0.50 | −0.12, +0.15 | −0.22, +0.42 | 0.55 | −0.13, +0.19 | −0.27, +0.42 | |
| or | (0.5 → 1.5) | 1.10 | −0.13, +0.21 | −0.24, +0.75 | 1.02 | −0.13, +0.13 | −0.29, +0.46 |
| R. Pechetti et al. [108] | F547M | JAM | JAM | ||||
| (0 → 8) | 6.67 | −0.09, +0.08 | −0.32, +0.18 | 6.72 | −0.15, +0.12 | −0.68, +0.26 | |
| () | (0.1 → 3) | 1.84 | −0.06, +0.07 | −0.15, +0.20 | 1.80 | −0.10, +0.09 | −0.24, +0.30 |
| (0.05 → 0.99) | 0.81 | −0.14, +0.13 | −0.34, +0.21 | 0.82 | −0.19, +0.14 | −0.44, +0.21 | |
| or | (0.5 → 1.5) | 0.98 | −0.09, +0.15 | −0.23, +0.75 | 1.00 | −0.18, +0.27 | −0.38, +0.78 |
Table 4. JAM models best-fitting parameters and formal uncertainties.
Dynamical Modelling
Jeans Anisotropic Models
We modeled the observed (and ) maps using both the cylindrically aligned (JAM; M. Cappellari [15]) and spherically aligned (JAM; M. Cappellari [17]) implementations of JAM, each assuming an axisymmetric potential, to assess how the assumed velocity-ellipsoid alignment affects the dynamical determination in NGC 5055.
In JAM, the velocity ellipsoid is aligned with cylindrical coordinates , with anisotropy parameterized as , where . In JAM, the ellipsoid is instead aligned with spherical coordinates , with anisotropy parameterized as , where . In both cases, all three velocity-dispersion components are permitted to differ. Models were computed using the jam_axi_proj routine in JAMPy (M. Cappellari [17]), with align='cyl' and align='sph' for the respective alignments, yielding the projected second-moment map , in which, the intrinsically projected second moments of and for JAM and of and for JAM. This approximation is purely a numerical convenience that enables fast model evaluation predictions for JAM, whereas JAM always computes the BH gravitational potential exactly.
Both JAM and JAM fit for (sampled logarithmically) and, linearly, the stellar mass-to-light ratio ( or ), the inclination (, internally converted to the minimum intrinsic axial ratio ), and the orbital anisotropy ( or ), which was converted into the equivalent dispersion ratios and for display purposes, respectively. The initial guesses and uniform priors over the search ranges for these four parameters are listed in Table 4. Given the limited radial extent of the OSIRIS FOV (, or pc at Mpc) and the absence of a reliable color map — the only available HST bands covering the nucleus are F606W and F547M, separated by only Å, too small a baseline for a robust constraint on spatial variations — we thus adopt spatially constant anisotropy and in both models. This assumption is further supported by the DESI nuclear spectrum, which is well fit by a single old ( Gyr), metal-rich () stellar population with no evidence for multiple components within the 1.1″5 fiber (§ 6.2). The same FOV also precludes a meaningful constraint on radially varying anisotropy, but see § 6.3 for the impact of testing varying anisotropy profiles ( or ) on the determination.
In total, we perform four JAM model runs spanning two choices of velocity-ellipsoid alignment: cylindrical (JAM) or spherical (JAM), and two stellar mass models: our three-component F606W parametric fit and the R. Pechetti et al. [108] F547M model (Table 3), to bracket the dominant sources of systematic uncertainty in our dynamical determination.
MCMC Fitting and Uncertainty Estimation
We sampled the JAM parameter space using a Markov chain Monte Carlo (MCMC) approach, employing the adaptive Metropolis algorithm (H. Haario et al. [56]) as implemented in the ADAMET package13 (M. Cappellari et al. [24]). Each chain comprised steps, with the first 20% discarded as burn-in. Best-fit parameters correspond to the maximum-likelihood solution, and uncertainties are quoted as (16th–84th percentile) and (0.14th–99.86th percentile) confidence levels (CLs) from the posterior distributions.
The large number of high-precision OSIRIS kinematic bins introduces two challenges for the JAM modeling. First, formal statistical uncertainties on the fitted parameters can be unrealistically small when unmodeled systematics are present. Second, the statistic can be disproportionately driven by the many data points at large radii, potentially biasing away from the kinematics within the SOI where the BH gravitational potential dominates.
To account for potential systematic errors in the measurements due to those two issues, we adopt the -inflation scheme of R. C. E. van den Bosch & G. van de Ven (2009), which in a Bayesian context is equivalent to scaling the kinematic uncertainties by , where is the number of kinematic bins [87 Section 6.1]. Rather than applying this inflation uniformly, we apply it selectively: formal uncertainties are preserved for all bins inside the SOI (), where the gravitational influence of the SMBH dominates the kinematics, while the uncertainties of the bins outside this radius are inflated by . This strategy ensures that is primarily constrained by the central kinematics while still propagating realistic systematic uncertainties at larger radii.
This approach is now standard practice in both molecular gas- (e.g., [120, 121, 107, 91, 97, 98, 99, 105 e]) and stellar-based dynamical measurements (e.g., [36, 125, 103 c] (R. C. E. van den Bosch & G. van de Ven 2009)).
To test whether depends on the choice of inflation radius, we ran three additional JAM fits: inflating at , at , and excluding all bins at with no inflation applied (69 of 82 bins retained). In each case, the best-fit from all four JAM configurations (JAM/JAM × two MGE models) is consistent with the original within the uncertainties, and the median across the four models shifts by less than 1%: from (original) to , , and , respectively. The BH mass is insensitive to the specific inflation radius adopted.
Results and Discussion
Supermassive Black Hole Mass
The best-fit parameters from all four JAM models are summarized in Table 4. The two JAM models yield and for the R. Pechetti et al. [108] and our refit D. B. Fisher & N. Drory [41] MGEs, respectively, while the two JAM models yield and . All four measurements are consistent within their uncertainties, confirming that is insensitive to both the assumed velocity-ellipsoid alignment and the choice of photometric model. However, the JAM models systematically return slightly larger values and broader uncertainties than their JAM counterparts, consistent with the greater degeneracy between the radial anisotropy parameter and in spherically aligned models (e.g., [17, 125, 104]).
The sphere of influence, R_{\mathrm{SOI}} = GM_{\mathrm{BH}}/\sigma_{\mathrm{e}}^{2} \approx0\farcs048, is only marginally resolved with respect to the PSF FWHM of 0\farcs045 (Figure 1), falling below the factor-of-two criterion commonly adopted as a threshold for reliable measurements (e.g., L. Ferrarese & H. Ford [39]). We note, however, that the JAM models are fitted simultaneously to all 82 spatial bins across the OSIRIS FOV; the constraint on arises from the overall shape of the 2D stellar velocity dispersion profile; specifically, the central rise that models with significantly lower fail to reproduce, rather than from the spatial resolution of the SOI alone. The robustness of this constraint is further demonstrated by seven independent systematic tests in § 6.3, all of which yield values well within of the fiducial result.
The larger uncertainties from our refit D. B. Fisher & N. Drory [41] MGE relative to R. Pechetti et al. [108] are primarily driven by its lower and less well-constrained –, which introduces greater degeneracy between inclination, , and . This corresponds to , in closer agreement with the large-scale H I disk inclination (; G. Battaglia et al. [8]), as expected given that the D. B. Fisher & N. Drory profile extends to \approx660\arcsec and is thus sensitive to the large-scale disk geometry. In contrast, the R. Pechetti et al. [108] MGE yields –, exceeding the observed galaxy axial ratio (), indicating that the JAM model finds an intrinsically round nuclear stellar distribution within the inner \approx10\arcsec, consistent with a compact, nearly circular nuclear component embedded within the more flattened outer disk, and precluding a meaningful inclination estimate from this model alone.
Similar to the derived , all models also consistently recover nearly isotropic stellar orbits (), with – for JAM and – for JAM. This is consistent with results from JWST/NIRSpec (D. D. Nguyen et al. [103]) and VLT/MUSE (D. A. Simon et al. [119]) stellar dynamical studies of similar systems, and with the mild-to-near-isotropic anisotropy commonly found in rotating galaxies from JAM modeling (M. Cappellari et al. [23]; M. Cappellari [15]).
The dynamical values are consistent between the two velocity-ellipsoid alignments, with – for the R. Pechetti et al. [108] MGE and – for our refit D. B. Fisher & N. Drory [41] MGE, the latter being somewhat lower as expected from the redder passband of F606W relative to F547M. The independent stellar population analysis of the DESI nuclear spectrum, yielding for a Gyr, supersolar-metallicity population, is in excellent agreement with our dynamical and provides stronger external validation (see Figure 8).
Figure 8. DESI DR1 nuclear spectrum of NGC 5055 extracted within a 1.′′5-diameter fiber aperture, fitted with PEGASE.HR stellar population models (D. Le Borgne et al. [73]) using pPXF. The observed spectrum (black) is overlaid with the best-fit stellar population model (red); fitting residuals are shown in blue; emission lines in green and the NaD absorption in violet. Inset: confidence contours in the stellar age–metallicity plane, shown at 1 (black), 2, 3, 5 (blue) levels. The well-constrained minimum at Gyr and [Fe/H] confirms a moderately old, highly super-solar nuclear stellar population.
Figure 5 shows the OSIRIS velocity dispersion () profiles of NGC 5055 extracted along the major and minor axes, with error bars inflated as described in § 5.2, overlaid with the best-fit JAM and JAM models (corresponding to the two adopted MGE stellar mass models for NGC 5055). All four models provide a good match to the observed profiles along both axes across the OSIRIS FOV.
Figure 6 shows the 2D posterior distributions from the ADAMET MCMC optimization for all four JAM models. Point colors indicate relative likelihood, with white denoting the maximum-likelihood solution and the CL, and black the boundary. The 1D marginalized histograms along the diagonal yield the best-fit parameter values and uncertainties, incorporating both stellar kinematic and statistical errors. The inset velocity dispersion maps in the upper-right corner of each panel compare the observed and best-fit model, with the map plotted against the world coordinate system (WCS) orientation: north is up and east to the left. All four models reproduce the observed kinematics to within relative residuals across the OSIRIS FOV.
Figure 6. Posterior distributions from the AdaMet MCMC optimization of the JAM models for the OSIRIS nuclear stellar kinematics of NGC 5055, shown after burn-in. The corner plot displays the 2D joint and 1D marginalized posterior distributions of , , , and . Solid and dashed vertical lines mark the best-fit values and uncertainties, respectively. Inset panels: observed (top) and best-fit model (bottom) maps on a consistent velocity scale, demonstrating the high fidelity of the fit across the OSIRIS FOV. Top-row plots: Results for (left) and (right) with the R. Pechetti et al. [108] MGE photometric (mass) model. Bottom-row plots: Same as the top-row plots, but with the D. B. Fisher & N. Drory [41] MGE photometric (mass) model.
In the JAM models, and exhibit a negative covariance: a higher assigns more mass to the stellar component, reducing the BH contribution needed to reproduce the observed peak. This classic trade-off is a well-known systematic in stellar dynamical modeling (M. Cappellari [15]; S. Thater et al. [125]), and vice versa. In contrast, the JAM models exhibit a positive – covariance, a consequence of the spherically aligned velocity ellipsoid: the radial anisotropy naturally promoted by this geometry inflates the projected line-of-sight such that both and must increase together to reproduce the observed kinematics, rather than trading off against each other.
A positive covariance between the anisotropy parameter ( for JAM and for JAM) and both and is consistently observed across all four models, reflecting the classical mass–anisotropy degeneracy (J. Binney & G. A. Mamon [11]; O. E. Gerhard [43]). More isotropic orbits produce a weaker central peak, requiring a higher to reproduce the observed kinematics; simultaneously, more radially biased orbits inflate the line-of-sight , which can equivalently be compensated by increasing , making anisotropy, , and difficult to disentangle from kinematics alone. This degeneracy represents an additional source of systematic uncertainty in our determination, partially mitigated by the fact that all four models consistently prefer nearly isotropic orbits (), limiting the range of allowed anisotropy and thus the associated uncertainty on . Additionally, exhibits a consistent negative covariance with : a rounder (more face-on) distribution produces weaker line-of-sight projection of the stellar velocity field, reducing the stellar mass contribution needed to reproduce the observed kinematics and thus requiring a lower .
Figure 7 shows our four measurements and their 1 uncertainties from the four conventional JAM models, providing a summary of the most precise constraint for NGC 5055 based on the Keck/OSIRIS stellar kinematics. We derive a robust SMBH estimate by adopting the median and its 1 statistic-bootstrapping CL across the four models, yielding .
Figure 7. Summary of measurements for NGC 5055 from four JAM models applied to the Keck/OSIRIS stellar kinematics, spanning two velocity-ellipsoid alignments (JAM, JAM) and two photometric models (F547M, F606W). Individual measurements with their uncertainties are shown as points with error bars. The vertical dashed line and shaded region indicate the adopted and its total quadrature (statistical and systematic) uncertainty, respectively.
To construct a comprehensive uncertainty budget, we combine in quadrature two contributions: (1) the statistical uncertainty from the MCMC posterior of each individual JAM model, which incorporates the propagated stellar kinematic measurement errors (including the selective error inflation described in § 5.2); and (2) the systematic uncertainty estimated from the model-to-model spread across the four JAM runs, reflecting variations in the assumed velocity-ellipsoid alignment and stellar mass model. Following common practice in dynamical determinations, the uncertainty due to the assumed distance is excluded from this budget, as scales linearly with distance and this contribution is treated separately (§ 6.3). This yields our final adopted mass of for NGC 5055.
Our dynamically measured is 30% lower than the value predicted by the N. Sahu et al. [113] – relation (§ 1.1), which is not surprising given that scaling relations carry intrinsic scatter of – dex and are calibrated predominantly on ETGs with classical bulges, while the morphological classification of NGC 5055’s central structure remains uncertain (§ 4). This stellar-dynamical yields ( pc), comparable to the Keck/OSIRIS PSF FWHM, confirming that the gravitational influence of the SMBH is detected within the central resolution element of our data.
Independent Stellar Population Validation from DESI Spectroscopy
We obtained an independent constraint on the stellar of the NGC 5055 nucleus from a high S/N spectrum provided by the Dark Energy Spectroscopic Instrument (DESI; DESI Collaboration et al. [33]) Data Release 1 (DR1), obtained within a circular fiber aperture of diameter 1.′′5 centered on the nucleus. The spectrum was fitted with the PEGASE.HR stellar population models (D. Le Borgne et al. [73]) adopting the Kroupa initial mass function (IMF; P. Kroupa [69]), using the updated NBURSTS full spectrum fitting (I. V. Chilingarian et al. 2007; I. Chilingarian et al. 2007; N. G. Pulatova et al. [111]). The model included a stellar population template, emission lines (H–, [O II] 4959/5007 Å, [O I] 6300/6364 Å, [N II] 6548/6584 Å, [S II] 6717/6731 Å) with the flux ratios of forbidden oxygen and nitrogen doublets fixed to the values from the NIST atomic physics database, and additional absorption components in the neutral sodium doublet (NaD). All three components have independent kinematics (, ).
We note that the more commonly used XSL (K. Verro et al. [134]) and eMILES (A. Vazdekis et al. [132]) stellar population template grids leave substantial spectral residuals around Å, indicative of the spectrum reaching the upper metallicity boundary of those grids ([Fe/H] = +0.2 dex for XSL and +0.4 for MILLES). PEGASE.HR, while lacking
coverage redward of Å, yields a significantly better fitting quality over the range 3900–6800 Å and extends to higher metallicities, making it the preferred choice for this nucleus with the super-solar metallicity. explains why we do not observe substantial extinction in the optical images of NGC 5055 in its nuclear region.
The best-fit stellar population parameters are: a luminosity-weighted age of Gyr and a metallicity of [Fe/H] = , yielding a -band stellar mass-to-light ratio of M/L. This is in excellent agreement with our dynamical values of – and – M/L from the JAM models (Table 4), providing strong independent validation of the stellar recovered from our dynamical modeling.
The DESI spectrum also yields three distinct velocity dispersion measurements from different kinematic tracers: the stellar continuum gives km s, the emission lines give km s, and the NaD absorption gives km s. The stellar km s exceeds the OSIRIS global dispersion of km s (§ 3.2), which is trivially explained by the beam-smearing effect: the galaxy rotation within the DESI aperture broadens both absorption and emission lines. The significantly lower dispersions of the emission lines and NaD feature relative to the stellar continuum indicate that these two tracers likely originate in the thin disk component and do not participate in the bulk stellar kinematics. The NaD absorption, in particular, arises primarily from neutral atomic gas along the line of sight (e.g., [110]) rather than from the stellar photosphere, and is therefore not a reliable tracer of stellar dynamics in this system. The ratio of equivalent widths of the NaD lines in the doublet ( and Å) indicates partial saturation of the 5889 Å line because the ratio is less than 2, the value expected in the non-saturated case. The measured equivalent width of the fainter 5896 Å line translates to the column density cm (see, e.g., [35]) and mag and
Additional Uncertainty Budgets
Distances: The distance to NGC 5055 is well constrained with Mpc from the TRGB measurement of K. B. W. McQuinn et al. [84], which represents one of the most reliable distance indicators for nearby galaxies. Nevertheless, the NASA/IPAC NED lists 27 independent distance determinations spanning 4.4–10.9 Mpc, reflecting the range of methods and their associated systematics. Since , this distance uncertainty propagates directly and linearly into our black hole mass, yielding a systematic range of – M, which represents a dominant component of the total error budget.
Stellar Kinematic Measurements with PHOENIX Synthetic Library: As an additional systematic check, we re-extracted the stellar kinematics of NGC 5055 using the PHOENIX synthetic stellar library ([64]) in place of the fiducial XSL library. The two sets of kinematics agree to within 2% (§ 3.3), and the resulting differences in and are negligible.
Fixed : We also ran four JAM models with fixed at 1.9, the value measured from the DESI spectrum (§ 6.2). The median across the four models shifts from M (free ) to M (3.5%), within of the original. is insensitive to whether is held fixed at the DESI value or left as a free parameter.
Varying Anisotropy Profile: We replace the constant anisotropy with a logistic profile ([119, 103 c]) for the JAM model and a radially varying anisotropy profile (Eq. 37 of M. Cappellari [20]) for the JAM model. Across all four JAM models, shifts by , well within the uncertainties.

Figure 11. – scaling diagram adapted from A. W. Graham [47] with updates from A. W. Graham et al. [52]. Massive merger-built elliptical (and ellicular) galaxies with are excluded given that their NSCs, if they still possess one, are expected to have been eroded by the merging process. Flattened nuclear disks can be many tens of parsecs in size (e.g., M. Balcells et al. [6]), and are not expected to follow the relation shown here. The two systems with the highest black mass belong to NGC 4026 and NGC 4697.
The MBH–Galaxy Scaling Diagrams
While NGC 5055 contains a prominent excess over the inner few arcseconds, brightening the surface brightness by more than mag arcsec, we regard the galaxy as bulgeless. As shown in Figure 9, traversing the light profile from large to small radii, the disk brightens at an accelerating pace (anti-truncates) upon approaching the outer domain () of the weak grand design spiral (§ 1.1), and anti-truncates again at upon approaching the tightly-wound ring-like spiral arms visible in the middle panel of Figure 9 (see also M. D. Thornley & L. G. Mundy [126]). The central excess is therefore attributed to the superposition of these nested disk substructures, a circumnuclear stellar disk, nuclear spiral arms, and the inner anti-truncated disk, rather than a classical or pseudobulge component. NGC 5055 thus joins a small sample of bulgeless spirals with direct dynamical determinations, including NGC 2748 (J. W. Atkinson et al. [5]) and NGC 4395 (M. den Brok et al. [32]), extending the BH mass census into a regime where scaling relations remain poorly constrained and each new measurement serves as a valuable anchor.

Figure 9. Left: SDSS DR8 image of NGC 5055 ([2])14. Middle: HST image (Credit: ESA/Hubble & NASA). Right: Variant of Figure 4 using a linear radial scale and a two-component fit. An inner Sérsic component (dashed red line) again approximates the complexities of the central structure while a revised anti-truncated exponential disk (solid blue line) reveals a disk displaying a double anti-truncated behaviour.
Using a (Galactic extinction)-corrected (E. F. Schlafly & D. P. Finkbeiner [115]) colour of 0.704 (G. de Vaucouleurs et al. [31]) obtained via the NASA Extragalactic Database (NED)15, gives a 3.6 m diet-Salpeter (E. F. Bell & R. S. de Jong [10]) mass-to-light ratio of 0.46 (A. W. Graham & N. Sahu [54, 55]). Applying this to the 3.6 m galaxy apparent magnitude of 7.89 AB-mag (B. L. Davis et al. [30]), and using a distance modulus of 29.74 mag, and an absolute magnitude for the Sun of 6.02 mag (AB), gives a stellar mass dex. This agrees with the value of 10.78 dex reported by T. C. Licquia & J. A. Newman [75] and G. Morales et al. [88], and dex reported by B. L. Davis et al. [30].
Given that the intermediate-scale Sérsic component that was fit in Figure 4 does not describe a bulge, but rather the inner part of the disk, we show NGC 5055 in the – diagram (Figure 10), but not the – diagram. NGC 5055 can be seen to reside within the cloud of spiral galaxies in both this diagram and the – diagram (Figure 10), revealing nothing unusual with the black hole mass derived herein. Notably, NGC 5055 lies dex below the scaling relations defined by elliptical galaxies and AGN hosts, placing it in a similar position to the Milky Way, whose central SMBH has also been noted as undermassive relative to these relations [139, 74]. This offset likely reflects a morphological dependence of the – relation, with late-type galaxies and bulgeless systems systematically falling below the locus defined by early-type hosts.
Figure 10. The masses are from A. W. Graham [50] and the velocity dispersions from N. Sahu et al. [113]. The equations in the left-hand panel are from A. W. Graham [51], and pertain to elliptical (E) galaxies (including brightest cluster galaxies (BCGs): steep dashed line), lenticular (S0) galaxies (solid line), and S0 galaxies with no visible signs of dust (shallow dashed line). The named galaxies were excluded from the regression analyses. The equations in the right-hand panel are from A. W. Graham [49] and pertain to galaxies built from major mergers (E, BCG, and dust-rich S0 galaxies: solid line), BCGs alone (parallel dashed line), and spiral (S) galaxies. With the exception of a few suspected faded S galaxies in the lower-right, the dust-poor S0 galaxies (dust = N and n) tend to reside to the low , and thus high , side of the relation defined by the local galaxies containing SMBHs, as noted in A. W. Graham et al. [52]. NGC 5055 falls dex below the relations defined by elliptical galaxies and AGN hosts, reminiscent of the Milky Way’s SMBH which has similarly been noted as undermassive relative to these relations ([139, 74]).
Despite lacking a stellar bar and the associated (peanut shell)-shaped structure of a buckled bar, the Sbc [31] galaxy NGC 5055 is remarkably similar in mass to the Milky Way (an SBbc galaxy), which has dex [74] and [28 cf., 4.56 \times 10^{6}\ M_{\odot}$ for NGC 5055). Furthermore, the mass of the nuclear star cluster in the Milky Way is $3.0 \times 10^{7}\ M_{\odot}$ {citep}][71], which also compares closely with that in NGC 5055 ($3.66 \times 10^{7}\ M_{\odot}$; {citep}R. Pechetti et al. 2020). As shown in {xref}Figure 11, NGC 5055, intriguingly, serves as yet another system following the $M_{\mathrm{BH}}$–$M_{\mathrm{NSC}}$ scaling relation {citep}[46], prior to the anticipated star cluster erosion in dry galaxy mergers {citep}[9]. The robustness of this relation is of critical importance for forecasting the rates of extreme-mass-ratio inspiral (EMRI) events, wherein compact stellar remnants (e.g., neutron stars and stellar-mass black holes) residing within the dense nuclear star cluster emit gravitational waves as they are captured by the central massive black hole {citep}[e.g. 3]`.
Conclusions
We have presented the first stellar-dynamical measurement of the SMBH mass in the nearby spiral galaxy NGC 5055 (M63) using LGS-AO IFU with Keck/OSIRIS. The unprecedented high-angular-resolution PSF FWHM of 0\farcs045, corresponding to 1.9 pc at Mpc) NIR observations resolve the 2D stellar kinematics that are robust against dust obscuration in the nucleus. Our main results are as follows:
High-resolution stellar nuclear kinematics. The 2D LOSVDs extracted from the CO bandheads () via pPXF reveal a velocity dispersion rising from at large radii to a central peak of within r \lesssim0\farcs1, providing unambiguous kinematic evidence for a central dark mass.
Robust black hole mass determination. We modeled the observed kinematics with an ensemble of four JAM models spanning two velocity-ellipsoid alignments (JAM and JAM) and two independent photometric decompositions. All variants yield consistent values within , and the median across models gives a fiducial black hole mass of , with the SOI (R_{\mathrm{SOI}} \approx0\farcs048, or pc) resolved by the Keck/OSIRIS PSF.
Stellar orbital structure. Both JAM and JAM consistently recover nearly isotropic stellar orbits in the nuclear region, with – () for JAM and – () for JAM, consistent with the near-isotropic orbital structure found in stellar dynamical studies of rotating LTGs (e.g., M. Cappellari [15]; D. D. Nguyen et al. [104]).
Photometric robustness. Stellar mass models from two independent HST photometric decompositions [108, 41], our refit) yield values consistent within 14%, confirming the robustness of our measurement against the choice of photometric model.
Image decomposition. Improvements of either two- or three-component parametric fit to the galaxy’s composite -band surface brightness profile disfavors the presence of a classical bulge component (Figure 4 and Figure 9, respectively), suggesting that NGC 5055 may be bulgeless, a morphological classification that, if confirmed, would place it among the rare spirals with a direct dynamical measurement in the absence of a classical bulge.
Black hole mass scaling relations. Our black hole mass measurement for NGC 5055 places it squarely within the cloud of data points for other spiral galaxies in the – and – diagrams (Figure 10). It also follows the – scaling relation, similar to our Milky Way, reinforcing the secular-evolution between the BH and the NSC in NGC 5055.
ACKNOWLEDGEMENTS
DDN acknowledges support from the ELT postdoctoral research fellowship at the Department of Astronomy, University of Michigan. RS acknowledges support from the INAF grant number 1.05.23.04.04. IC’s research is supported by the SAO Telescope Data Center. He also acknowledges support from the NASA ADAP-22-0102 and HST-GO-16739 grants.
Some of the data presented herein were obtained at Keck Observatory, which is a private 501(c)3 non-profit organization operated as a scientific partnership among the California Institute of Technology, the University of California, and the National Aeronautics and Space Administration. The Observatory was made possible by the generous financial support of the W. M. Keck Foundation.
Facility: Keck/OSIRIS. Keck Program ID: 2024B_W294. The authors wish to recognize and acknowledge the very significant cultural role and reverence that the summit of Maunakea has always had within the Native Hawaiian community. We are most fortunate to have the opportunity to conduct observations from this mountain.
Software: Python 3.12: [131], Matplotlib 3.6: [63], NumPy 1.22: [58], SciPy 1.3: [135], photutils 0.7: [13], AstroPy 5.1 [4], AdaMet 2.0 [24], JamPy 7.2 [17], MgeFit 5.0 [14], and Profiler [27].
References
- [1]Afanasiev, V. L., & Sil’chenko, O. K. 2002, A&A, 388, 461, doi: 10.1051/0004-6361:20020527
- [2]Aihara, H., Allende Prieto, C., An, D., et al. 2011, ApJS, 193, 29, doi: 10.1088/0067-0049/193/2/29DOI
- [3]Amaro-Seoane, P., Audley, H., Babak, S., et al. 2017, arXiv e-prints, arXiv:1702.00786, doi: 10.48550/arXiv.1702.00786
- [4]Astropy Collaboration, Price-Whelan, A. M., Lim, P. L., et al. 2022, ApJ, 935, 167, doi: 10.3847/1538-4357/ac7c74DOI
- [5]Atkinson, J. W., Collett, J. L., Marconi, A., et al. 2005, MNRAS, 359, 504, doi: 10.1111/j.1365-2966.2005.08904.xDOI
- [6]Balcells, M., Graham, A. W., & Peletier, R. F. 2007, ApJ, 665, 1084, doi: 10.1086/519752DOI
- [7]Barth, A. J., Ho, L. C., & Sargent, W. L. W. 2002, AJ, 124, 2607, doi: 10.1086/343840DOI
- [8]Battaglia, G., Fraternali, F., Oosterloo, T., & Sancisi, R. 2006, A&A, 447, 49, doi: 10.1051/0004-6361:20053017
- [9]Bekki, K., & Graham, A. W. 2010, ApJL, 714, L313, doi: 10.1088/2041-8205/714/2/L313DOI
- [10]Bell, E. F., & de Jong, R. S. 2001, ApJ, 550, 212, doi: 10.1086/319728DOI
- [11]Binney, J., & Mamon, G. A. 1982, MNRAS, 200, 361, doi: 10.1093/mnras/200.2.361DOI
- [12]Blais-Ouellette, S., Carignan, C., Amram, P., & Côté, S. 1999, AJ, 118, 2123, doi: 10.1086/301066
- [13]Bradley, L., Sipőcz, B., Robitaille, T., et al. 2024, 2.0.2 Zenodo, doi: 10.5281/zenodo.13989456
- [14]Cappellari, M. 2002, MNRAS, 333, 400, doi: 10.1046/j.1365-8711.2002.05412.xDOI
- [15]Cappellari, M. 2008, MNRAS, 390, 71, doi: 10.1111/j.1365-2966.2008.13754.xDOI
- [16]Cappellari, M. 2017, MNRAS, 466, 798, doi: 10.1093/mnras/stw3020DOI
- [17]Cappellari, M. 2020, MNRAS, 494, 4819, doi: 10.1093/mnras/staa959DOI
- [18]Cappellari, M. 2023, MNRAS, 526, 3273, doi: 10.1093/mnras/stad2597DOI
- [19]Cappellari, M. 2025, MNRAS, 544, 1432, doi: 10.1093/mnras/staf1726DOI
- [20]Cappellari, M. 2026, MNRAS, 549, stag420, doi: 10.1093/mnras/stag420DOI
- [21]Cappellari, M., & Emsellem, E. 2004, PASP, 116, 138, doi: 10.1086/381875DOI
- [22]Cappellari, M., Neumayer, N., Reunanen, J., et al. 2009, MNRAS, 394, 660, doi: 10.1111/j.1365-2966.2008.14377.xDOI
- [23]Cappellari, M., Emsellem, E., Bacon, R., et al. 2007, MNRAS, 379, 418, doi: 10.1111/j.1365-2966.2007.11963.xDOI
- [24]Cappellari, M., Scott, N., Alatalo, K., et al. 2013, MNRAS, 432, 1709, doi: 10.1093/mnras/stt562DOI
- [27]Ciambur, B. C. 2016, PASA, 33, e062, doi: 10.1017/pasa.2016.60DOI
- [28]Collaboration, T. E. H. T. 2022, The Astrophysical Journal Letters, 930
- [29]Davies, R. L., Efstathiou, G., Fall, S. M., Illingworth, G., & Schechter, P. L. 1983, ApJ, 266, 41, doi: 10.1086/160757DOI
- [30]Davis, B. L., Graham, A. W., & Cameron, E. 2018, ApJ, 869, 113, doi: 10.3847/1538-4357/aae820DOI
- [31]de Vaucouleurs, G., de Vaucouleurs, A., Corwin, Jr., H. G., et al. 1991, Third Reference Catalogue of Bright Galaxies, Vol. 1 (Springer, New York)
- [32]den Brok, M., Seth, A. C., Barth, A. J., et al. 2015, ApJ, 809, 101, doi: 10.1088/0004-637X/809/1/101DOI
- [33]DESI Collaboration, Abdul Karim, M., Adame, A. G., et al. 2026, AJ, 171, 285, doi: 10.3847/1538-3881/ae4c43
- [34]Di Matteo, T., Springel, V., & Hernquist, L. 2005, Nature, 433, 604, doi: 10.1038/nature03335DOI
- [35]Draine, B. T. 2011, Physics of the Interstellar and Intergalactic Medium
- [36]Drehmer, D. A., Storchi-Bergmann, T., Ferrari, F., Cappellari, M., & Riffel, R. A. 2015, MNRAS, 450, 128, doi: 10.1093/mnras/stv536DOI
- [37]Elmegreen, D. M., & Elmegreen, B. G. 1987, ApJ, 314, 3, doi: 10.1086/165034DOI
- [38]Emsellem, E., Monnet, G., & Bacon, R. 1994, A&A, 285, 723
- [39]Ferrarese, L., & Ford, H. 2005, Space Science Reviews, 116, 523, doi: 10.1007/s11214-005-3947-6DOI
- [40]Ferrarese, L., & Merritt, D. 2000, ApJL, 539, L9, doi: 10.1086/312838DOI
- [41]Fisher, D. B., & Drory, N. 2008, AJ, 136, 773, doi: 10.1088/0004-6256/136/2/773DOI
- [42]Gebhardt, K., Bender, R., Bower, G., et al. 2000, ApJL, 539, L13, doi: 10.1086/312840DOI
- [43]Gerhard, O. E. 1993, MNRAS, 265, 213, doi: 10.1093/mnras/265.1.213DOI
- [44]Graham, A. W. 2008, PASA, 25, 167, doi: 10.1071/AS08013DOI
- [45]Graham, A. W. 2012, ApJ, 746, 113, doi: 10.1088/0004-637X/746/1/113DOI
- [46]Graham, A. W. 2016, in Astrophysics and Space Science Library, Vol. 418, Galactic Bulges, ed. E. Laurikainen, R. Peletier, & D. Gadotti, 263, doi: 10.1007/978-3-319-19378-6_11
- [47]Graham, A. W. 2020, MNRAS, 492, 3263, doi: 10.1093/mnras/stz3547DOI
- [48]Graham, A. W. 2023a, MNRAS, 518, 6293, doi: 10.1093/mnras/stac3173DOI
- [49]Graham, A. W. 2023b, MNRAS, 522, 3588, doi: 10.1093/mnras/stad1124DOI
- [50]Graham, A. W. 2026a, MNRAS, doi: 10.1093/mnras/stag782DOI
- [51]Graham, A. W. 2026b, MNRAS, submitted
- [52]Graham, A. W., Chilingarian, I., Nguyen, D. D., et al. 2025, PASA, 42, e068, doi: 10.1017/pasa.2025.10035DOI
- [53]Graham, A. W., Erwin, P., Caon, N., & Trujillo, I. 2001, ApJL, 563, L11, doi: 10.1086/338500DOI
- [54]Graham, A. W., & Sahu, N. 2023, MNRAS, 518, 2177, doi: 10.1093/mnras/stac2019DOI
- [55]Graham, A. W., & Sahu, N. 2024, MNRAS, 530, 3429, doi: 10.1093/mnras/stac1079
- [56]Haario, H., Saksman, E., & Tamminen, J. 2001, Bernoulli, 7, 223
- [57]Haehnelt, M. G., Natarajan, P., & Rees, M. J. 1998, MNRAS, 300, 817, doi: 10.1046/j.1365-8711.1998.01951.xDOI
- [58]Harris, C. R., Millman, K. J., van der Walt, S. J., et al. 2020, Nature, 585, 357, doi: 10.1038/s41586-020-2649-2DOI
- [59]Heckman, T. M. 1980, A&A, 87, 152
- [60]Ho, L. C. 2008, ARA&A, 46, 475, doi: 10.1146/annurev.astro.45.051806.110546DOI
- [62]Hu, J. 2008, MNRAS, 386, 2242, doi: 10.1111/j.1365-2966.2008.13195.xDOI
- [63]Hunter, J. D. 2007, Computing In Science & Engineering, 9, 90, doi: 10.1109/MCSE.2007.55DOI
- [64]Husser, T.-O., Wende-von Berg, S., Dreizler, S., et al. 2013, A&A, 553, A6, doi: 10.1051/0004-6361/201219058DOI
- [65]Jiménez-Vicente, J., & Aguerri, J. A. L. 2017, MNRAS, 469, 3564, doi: 10.1093/mnras/stx1005
- [66]King, A. 2003, ApJL, 596, L27, doi: 10.1086/379143DOI
- [67]Kormendy, J., & Ho, L. C. 2013, ARA&A, 51, 511, doi: 10.1146/annurev-astro-082708-101811DOI
- [68]Krajnović, D., Cappellari, M., de Zeeuw, P. T., & Copin, Y. 2006, MNRAS, 366, 787, doi: 10.1111/j.1365-2966.2005.09902.xDOI
- [69]Kroupa, P. 2001, MNRAS, 322, 231, doi: 10.1046/j.1365-8711.2001.04022.xDOI
- [70]Larkin, J., Barczyns, M., Krabbe, A., et al. 2006, New Astron Rev, 50, 362, doi: 10.1016/j.newar.2006.02.005
- [71]Launhardt, R., Zylka, R., & Mezger, P. G. 2002, A&A, 384, 112, doi: 10.1051/0004-6361:20020017DOI
- [73]Le Borgne, D., Rocca-Volmerange, B., Prugniel, P., et al. 2004, A&A, 425, 881, doi: 10.1051/0004-6361:200400044DOI
- [74]Lian, J., Wang, T., Feng, Q., Huang, Y., & Guo, H. 2025, ApJL, 990, L37, doi: 10.3847/2041-8213/adfc73DOI
- [75]Licquia, T. C., & Newman, J. A. 2015, ApJ, 806, 96, doi: 10.1088/0004-637X/806/1/96DOI
- [76]Magorrian, J., Tremaine, S., Richstone, D., et al. 1998, AJ, 115, 2285, doi: 10.1086/300535
- [77]Maoz, D., Filippenko, A. V., Ho, L. C., et al. 1998, ApJS, 114, 73, doi: 10.1086/313066
- [78]Maoz, D., Filippenko, A. V., Ho, L. C., et al. 1995, ApJ, 440, 91, doi: 10.1086/175250
- [79]Marconi, A., & Hunt, L. K. 2003, ApJL, 589, L21, doi: 10.1086/375804DOI
- [80]Marconi, A., Schreier, E. J., Koekemoer, A., et al. 2000, ApJ, 528, 276, doi: 10.1086/308168DOI
- [81]Martínez-Delgado, D., Gabany, R. J., Crawford, K., et al. 2010, AJ, 140, 962, doi: 10.1088/0004-6256/140/4/962
- [82]Masi, G. 2014, in European Planetary Science Congress, Vol. 9, EPSC2014–740
- [83]McLure, R. J., & Dunlop, J. S. 2002, MNRAS, 331, 795, doi: 10.1046/j.1365-8711.2002.05236.xDOI
- [84]McQuinn, K. B. W., Skillman, E. D., Dolphin, A. E., Berg, D., & Kennicutt, R. 2017, AJ, 154, 51, doi: 10.3847/1538-3881/aa7aadDOI
- [87]Mitzkus, M., Cappellari, M., & Walcher, C. J. 2017, MNRAS, 464, 4789, doi: 10.1093/mnras/stx2677
- [88]Morales, G., Martínez-Delgado, D., Grebel, E. K., et al. 2018, A&A, 614, A143, doi: 10.1051/0004-6361/201732271DOI
- [91]Ngo, H. N., Nguyen, D. D., Le, T. Q. T., et al. 2025b, ApJ, 992, 211, doi: 10.3847/1538-4357/ae0455DOI
- [92]Nguyen, D. D., Cappellari, M., & Pereira-Santaella, M. 2023, MNRAS, 526, 3548, doi: 10.1093/mnras/stad2860DOI
- [93]Nguyen, D. D., Seth, A. C., Reines, A. E., et al. 2014, ApJ, 794, 34, doi: 10.1088/0004-637X/794/1/34DOI
- [94]Nguyen, D. D., Seth, A. C., den Brok, M., et al. 2017, ApJ, 836, 237, doi: 10.3847/1538-4357/aa5cb4DOI
- [95]Nguyen, D. D., Seth, A. C., Neumayer, N., et al. 2018, ApJ, 858, 118, doi: 10.3847/1538-4357/aabe28DOI
- [96]Nguyen, D. D., Seth, A. C., Neumayer, N., et al. 2019, ApJ, 872, 104, doi: 10.3847/1538-4357/aafe7aDOI
- [97]Nguyen, D. D., den Brok, M., Seth, A. C., et al. 2020, ApJ, 892, 68, doi: 10.3847/1538-4357/ab77aaDOI
- [98]Nguyen, D. D., Nguyen, T. H., Thater, S., et al. 2021, MNRAS, 504, 4123, doi: 10.1093/mnras/stab1002DOI
- [99]Nguyen, D. D., Bureau, M., Thater, S., et al. 2022, MNRAS, 509, 2920, doi: 10.1093/mnras/stab3016DOI
- [100]Nguyen, D. D., Ngo, H. N., Le, T. Q. T., et al. 2025a, A&A, 698, L9, doi: 10.1051/0004-6361/202554672DOI
- [102]Nguyen, D. D., Cappellari, M., Le, T. Q. T., et al. 2026a, MNRAS, 546, stag238, doi: 10.1093/mnras/stag238DOI
- [103]Nguyen, D. D., Ngo, H. N., Cappellari, M., et al. 2026b, ApJ, 999, 97, doi: 10.3847/1538-4357/ae32f2DOI
- [104]Nguyen, D. D., Le, T. N., Cappellari, M., et al. 2026c, ApJ, 1003, 98, doi: 10.3847/1538-4357/ae6460
- [105]Nguyen, D. D., Nguyen, L. Q. T., Gallo, E., et al. 2026d, ApJ, 1005, 94, doi: 10.3847/1538-4357/ae771cDOI
- [107]North, E. V., Davis, T. A., Bureau, M., et al. 2019, MNRAS, 490, 319, doi: 10.1093/mnras/stz2598DOI
- [108]Pechetti, R., Seth, A., Neumayer, N., et al. 2020, ApJ, 900, 32, doi: 10.3847/1538-4357/abaa7
- [109]Perez, I., Beckman, J. E., Rozas, M., Ryder, S., & Knapen, J. H. 1995, AJ, 109, 140, doi: 10.1086/117270
- [110]Poznanski, D., Prochaska, J. X., & Bloom, J. S. 2012, MNRAS, 426, 1465, doi: 10.1111/j.1365-2966.2012.21796.xDOI
- [111]Pulatova, N. G., Rubtsov, E., Chilingarian, I. V., et al. 2025, A&A, 702, A67, doi: 10.1051/0004-6361/202555117DOI
- [112]Saglia, R. P., Opitsch, M., Erwin, P., et al. 2016, ApJ, 818, 47, doi: 10.3847/0004-637X/818/1/47DOI
- [113]Sahu, N., Graham, A. W., & Davis, B. L. 2019a, ApJ, 887, 10, doi: 10.3847/1538-4357/ab50b7DOI
- [115]Schlafly, E. F., & Finkbeiner, D. P. 2011, ApJ, 737, 103, doi: 10.1088/0004-637X/737/2/103DOI
- [116]Silge, J. D., & Gebhardt, K. 2003, AJ, 125, 2809, doi: 10.1086/375324DOI
- [117]Silk, J., Begelman, M. C., Norman, C., Nusser, A., & Wyse, R. F. G. 2024, ApJL, 961, L39, doi: 10.3847/2041-8213/ad1bf0DOI
- [118]Silk, J., & Rees, M. J. 1998, A&A, 331, L1, doi: 10.48550/arXiv.astro-ph/9801013DOI
- [119]Simon, D. A., Cappellari, M., & Hartke, J. 2024, MNRAS, 527, 2341, doi: 10.1093/mnras/stad3309DOI
- [120]Smith, M. D., Bureau, M., Davis, T. A., et al. 2019, MNRAS, 485, 4359, doi: 10.1093/mnras/stz625DOI
- [121]Smith, M. D., Bureau, M., Davis, T. A., et al. 2021, MNRAS, 503, 5984, doi: 10.1093/mnras/stab791DOI
- [122]Springob, C. M., Haynes, M. P., Giovanielli, R., & Kent, B. R. 2005, ApJS, 160, 149, doi: 10.1086/431550DOI
- [123]Thater, S., Krajnović, D., Cappellari, M., et al. 2019, A&A, 625, A62, doi: 10.1051/0004-6361/201834808DOI
- [124]Thater, S., Lyubenova, M., Fahrion, K., et al. 2023, A&A, 675, A18, doi: 10.1051/0004-6361/202245362DOI
- [125]Thater, S., Krajnović, D., Weijmans, A., et al. 2022, MNRAS, 509, 5416, doi: 10.1093/mnras/stab3210DOI
- [126]Thornley, M. D., & Mundy, L. G. 1997, ApJ, 484, 202, doi: 10.1086/304306DOI
- [127]Toloba, E., Guhathakurta, P., Peletier, R. F., et al. 2014, ApJS, 215, 17, doi: 10.1088/0067-0049/215/2/17DOI
- [131]Van Rossum, G., & Drake, F. L. 2009, Python 3 Reference Manual (Scotts Valley, CA: CreateSpace)
- [132]Vazdekis, A., Koleva, M., Ricciardelli, E., Röck, B., & Falcón-Barroso, J. 2016, MNRAS, 463, 3409, doi: 10.1093/mnras/stw2231DOI
- [133]Verro, K., Trager, S. C., Peletier, R. F., et al. 2022a, A&A, 660, A34, doi: 10.1051/0004-6361/202142388DOI
- [134]Verro, K., Trager, S. C., Peletier, R. F., et al. 2022b, A&A, 661, A50, doi: 10.1051/0004-6361/202142387DOI
- [135]Virtanen, P., Gommers, R., Oliphant, T. E., et al. 2020, Nature Methods, 17, 261, doi: 10.1038/s41592-019-0686-2DOI
- [136]Wevers, B. M. H. R., van der Kruit, P. C., & Allen, R. J. 1986, A&AS, 66, 505
- [137]Wizinowich, P. L., Le Mignant, D., Bouchez, A. H., et al. 2006, PASA, 118, 297, doi: 10.1086/499290DOI
- [138]York, D. G., Adelman, J., Anderson, Jr., J. E., et al. 2000, AJ, 120, 1579, doi: 10.1086/301513DOI
- [139]Zou, F., Brandt, W. N., Gallo, E., et al. 2024, ApJ, 976, 6, doi: 10.3847/1538-4357/ad815dDOI






