Multi-band infrared imaging reveals dusty spiral arcs around the binary B[e] star 3 Puppis

TL;DR

The authors resolve the dusty circumbinary disc of the B[e] star 3 Puppis at mas scales using VLTI/MATISSE across L, M, and N bands, and introduce a rigorous MiRA-based statistical imaging workflow (with PYRA and MYTHRA) complemented by SPARCO to recover CE structures. They uncover a prominent SE elongated clump at ~16.7 mas (~10 au) and a fainter NW asymmetry, with the inner rim geometry consistent with prior K-band results but revealing broader, large-scale asymmetries. Geometric modelling and radiative-transfer comparisons show partial consistency with earlier AMBER/MIDI data, but the MATISSE images require an additional, spirally perturbed disc component; hydrodynamic analysis favours tidally induced spiral density waves driven by the central binary as the origin of the structures. The findings demonstrate that the CE around 3 Pup is dynamically driven by binary interaction rather than local gravitational instability or a third body, highlighting the role of binarity in shaping the dusty environments of B[e] stars and providing a framework for future high-angular-resolution studies.

Abstract

3 Puppis is the brightest known B[e] star. Recent work classifies this A-type object as a supergiant, yet the impact of its binarity on the circumstellar environment (CE) remains hard to characterize. To resolve its dusty region at 5-10 mas, we obtained mid-IR interferometric observations with VLTI/MATISSE over 3-12 μm. Because the (u,v) coverage supports imaging, we introduce a statistical interferometric-imaging workflow based on MiRA to generate averaged images: this systematic approach enables the selection of an optimal set of reconstructions, improving the robustness and fidelity of the recovered features. We also use SPARCO, an independent tool well suited to bright central objects embedded in fainter extended emission. Images from both tools in the L, M, and N bands agree and reveal an asymmetric, elongated feature ~17 mas (~10 au at 631 pc) southeast of the star with ~20% density contrast. A second northwest asymmetry and a skewed inner rim are detected. Simple geometric modelling, guided by the MATISSE images, constrains the morphology, location, and flux of the CE and its asymmetries. The images are consistent with earlier VLTI measurements but expose a more complex CE with large-scale clumps in the southeast and northwest parts of the disc. Hydrodynamic modelling indicates that tidal spiral-wake perturbations from the central binary, dynamically excited at Lindblad resonances in the circumbinary disc, best explain the radial extent and curvature of the elongated structures seen in all bands.

Figures (11)

• Figure 1: Coverage of the ($u,v$)-plane obtained from the VLTI/MATISSE observations of $3$ Puppis. The legend in the lower-right corner indicates the colours corresponding to the four AT configurations used: Small (blue), Medium (orange), Large (green), and Extended (red). A schematic map of the VLTI baselines, using the same colour code, is shown in the upper-right corner. East corresponds to increasing $x$-axis values, and North to increasing $y$-axis values.
• Figure 2: VLTI/MATISSE squared visibility data, denoted V$^2$, in logarithmic scale (top plot) and closure phase data, denoted CP, (bottom plot) of $3$ Puppis in the L- (blue), M- (green), and N- (red) bands, plotted as a function of spatial frequency B/$\lambda$. The vertical dashed lines indicate the approximate positions of the first visibility nulls in the three bands with the same colour code. The horizontal dashed line marks the visibility plateau at long baselines in the L-band, which reflects the relative flux contribution from the unresolved central source.
• Figure 3: Comparison between the interferometric observables extracted from the MYTHRA averaged images (in blue) and the VLTI/MATISSE data of $3$ Puppis (in green), shown respectively in the L-band (first row), in the M-band (second row), and in the N-band (third row). Left column: Squared visibility amplitudes as a function of spatial frequency in logarithmic scale. Central column: Closure phases as a function of spatial frequency, computed using the longest baseline. Right column: Final averaged images resulting from MYTHRA, normalized to unity (i.e. the sum over all pixels equals one), with a logarithmic colour scaling, and not convolved with the interferometric beam. The minimum value for each brightness scaling is set to half the standard deviation of the associated image to filter out the artefact produced by the MiRA algorithm.
• Figure 4: Final image reconstructions of $3$ Puppis in the mid-infrared using MiRA, MYTHRA, and SPARCO imaging tools. The field of view for each image is (60 mas $\times$ 60 mas) with North up and East left. The blue dashed ellipse indicates the best-fit inner rim of the dusty disc derived from VLTI/AMBER data, while the blue cross marks the southeastern elongated clump best-fit position, denoted $R_\mathrm{ref}$, from MATISSE L-band geometric modelling (at a distance of $16.71 \pm 0.03$ mas from the image centre). Brightness colour scale is normalised to peak intensity (i.e. maximum pixel value). K-band image: The left-most column shows the images obtained with VLTI/AMBER data by Millour+2011. In the first row the median MiRA image is displayed, and the last two rows present the identical convolved median MiRA image with the $\lambda/2B_\mathrm{max}$ PSF subtraction applied. L-M-N-bands images: The remaining three columns display the images obtained with VLTI/MATISSE. On the first row the resulting MYTHRA averaged image is shown. The second row shows the convolved image of MYTHRA with the $\lambda/2B_\mathrm{max}$ PSF subtraction applied. The last row gives the resulting SPARCO image, convolved with the interferometric beam.
• Figure 5: Spectral evolution of the fitted disc parameters from a chromatic two-component geometric model applied to $3$ Puppis N-band observations obtained with the VLTI/MIDI (in blue) and VLTI/MATISSE (in orange) instruments. Top plot: The full width at half maximum (FWHM) of the major-axis as a function of wavelength. Bottom plot: Elongation (or flattening) ratio as a function of wavelength.