The Supermassive Black Hole in the Nearby Spiral Galaxy M81: A Robust Mass from JWST/NIRSpec Stellar Dynamics

TL;DR

This paper presents the first robust stellar-dynamical measurement of the SMBH mass in the nearby spiral galaxy M81 using JWST/NIRSpec CO bandhead kinematics in the near-infrared. By employing Jeans Anisotropic Modeling within a Bayesian framework and exhaustively testing PSF, $M/L_J$, and orbital anisotropy, the authors derive $M_{ m BH} = (4.78^{+0.07}_{-0.10})\times10^{7} M_\odot$ with a well-resolved sphere of influence of about $0.5''$. The analysis demonstrates a clear mass–anisotropy degeneracy and rigorously quantifies systematic uncertainties, illustrating the impact of AGN contamination, distance, and kinematic templates. The result places M81 on established SMBH scaling relations for spirals, highlighting JWST’s critical role in refining SMBH demographics and providing a robust anchor for future studies.

Abstract

Despite its proximity, the mass of the supermassive black hole (SMBH) in the spiral galaxy M81 (NGC~3031) has remained uncertain, with previous dynamical measurements being unreliable. We present the first robust stellar-dynamical measurement of its mass using high-resolution, two-dimensional kinematics from JWST/NIRSpec observations of the central $3''\times3''$. By tracing stellar motions in the near-infrared, our data penetrate the obscuring nuclear dust and allow for the separation of stellar light from the non-thermal AGN continuum. We modeled the kinematics using JAM within a Bayesian framework, exploring a comprehensive suite of models that systematically account for uncertainties in the point-spread function, orbital anisotropy, and stellar mass-to-light ratio. This ensemble modeling approach demonstrates that a central dark mass unambiguously drives the central rise in velocity dispersion. The models yield a robust SMBH mass of $M_{\rm BH} = (4.78^{+0.07}_{-0.10})\times10^7$ M$_\odot$. This result resolves a long-standing uncertainty in the mass of M81's black hole and provides a crucial, reliable anchor point for SMBH-galaxy scaling relations.

Figures (14)

• Figure 1: Modeling of wiggles in the single-spaxel NIRSpec spectrum of M81. Top panel: Integrated spectrum (orange), single-spaxel spectrum (blue), and residual wiggles (gray). The red curve shows the best-fitting wiggle model. Bottom panel: Wiggle-corrected single-spaxel spectrum (dark blue) compared to the integrated spectrum (orange); residuals after correction are shown in gray. In all panels, red shaded regions mark emission lines excluded from the fit.
• Figure 2: Panel A: Logarithmically scaled intensity map of the NIRSpec G235H/F170LP data cube, collapsed along the spectral axis (excluding the detector gap at 2.41–2.49 ). The red circular annulus ($0\farcs4 < r < 1\farcs4$) indicates the region from which the global spectrum was extracted. Panel B: The observed global spectrum (black) overlaid with the best-fit stellar template (red) for M81. The fit residuals ( data - model) are shown in green and vertically offset by +0.62 to compress the $y$-axis range and better illustrate the stellar CO absorption bandheads. This same vertical offset is applied to all subsequent figures of this type.
• Figure 3: Radial spectral variation in M81 from the NIRSpec G235H/F170LP data cube. Left column: Observed spectra (black line) obtained by coadding spaxels within concentric annuli at radius $r$. The overplotted red line shows the pPXF fit, which combines the global stellar template (see left panel of Figure \ref{['fig:Intensity_map']} and Section \ref{['sec:kine']}), convolved with a Gaussian LOSVD and added by fourth-degree additive and multiplicative polynomials to account for the non-thermal nuclear continuum. Fit residuals ( data-model) are shown in green, with regions masked due to emission lines or artifacts indicated in gray. Right column: The convolved stellar template (red line) is compared to the observed spectrum after subtracting of the modeled non-thermal continuum (black line).
• Figure 4: The radial surface brightness profile $I(r)$, derived from individual Voronoi bins in the NIRSpec G235H/F170LP data cube of M81 (filled black circles), is compared to the estimated stellar light profile $\Gamma(r) = I(r)\gamma(r)$ (red open circles). The underlying stellar distribution is smooth and well approximated by a single-power-law (red line with pink region shows its $1\sigma$ uncertainty). The PSF–convolved central interpolation of this single power-law also reproduces the two innermost $\gamma$ measurements, which lie below the intrinsic (unconvolved) profile, indicating that their apparent decline is fully consistent with PSF effects. Gray points indicate pixels where $\gamma < 0.5$.
• Figure 5: Stellar kinematic measurements from the NIRSpec G235H/F170LP observations of NGC 3031 are shown. Panel A: The logarithmic collapsed intensity map (cf. left panel of Figure \ref{['fig:Intensity_map']}). Panel B: Example pPXF fit for a central-offset Voronoi bin unaffected by wiggles (location indicated in Panel A). The observed spectrum (black) displays the CO bandhead absorptions near 2.3 $\mu$m, with the best-fit XSL template overplotted in red. Fit residuals ( data-model) are plotted in green, and the vertical gray lines mark the wavelength range used for fitting across all bins. Panels C--F: 2D maps of the of the stellar rotation velocity ($V$), velocity dispersion ($\sigma$), root-mean-square velocity ($V_{\rm rms} = \sqrt{V^2 + \sigma^2}$), and stellar light contribution fraction ($\gamma$). The central white pixels are masked due to unreliable kinematics, as the AGN continuum contributes $\approx$66% of the total light in this region. White contours trace the intensity, decreasing by mag arcseconds$^{-1}$ steps from the center.