X TrA through the eyes of MATISSE: More evidence of clumpy molecular layers around C-type asymptotic giant branch stars

TL;DR

The study investigates wind formation around the carbon-rich AGB star X TrA by imaging its close environment (within a few stellar radii) using MATISSE in the L and N bands. Low-resolution interferometric imaging with image reconstruction (MiRA and SQUEEZE) reveals that the inner envelope is highly asymmetric and clumpy, especially in the 3.1 μm C2H2+HCN band, with angular diameters ranging from about 10 to 20 mas across wavelengths. The findings support a scenario where convection and pulsation drive localized, inhomogeneous gas and dust formation, influencing wind launching, and show no evidence for a binary companion within the sensitivity of the data. The results are consistent with similar carbon stars and demonstrate MATISSE’s capability to resolve the clumpy inner envelopes of C-type AGB stars, informing models of dust formation and mass loss in these objects.

Abstract

Aims. The goal of this study is to further the understanding of the wind formation mechanism in asymptotic giant branch (AGB) stars through the analysis of the close environment (within a few stellar radii) of the carbon star X TrA. Methods. X TrA was observed for the first time with the Mid-Infrared SpectroScopic Experiment instrument (MATISSE) in the L and N bands in low spectral resolution mode (R=30), and its close surroundings were mapped in specific wavelength ranges corresponding to specific molecules ($C_2H_2$ and HCN, at 3.1 and 3.8 $μ$m) and dust (amorphous carbon and, for example, Sic at 11.3 $μ$m), via image reconstruction techniques. Results. The angular diameter of the star ranges from 10 mas in the L band pseudo-continuum (3.5 $μ$m) to 20 mas at 3.1 and 11.3 $μ$m. The reconstructed images show some mild elongated features (along the east-west direction) and asymmetric protrusions, which are most evident around 3.1 $μ$m. Imaging results highlight the clumpy nature of the circumstellar environment, starting from the photospheric region up to more distant layers. Conclusions. The angular diameters found for X TrA in the image data are in agreement with previous photospheric diameter estimates (following VLTI/MIDI 8-13 $μ$m observations), and their wavelength dependence is similar to values found for other carbon stars observed with MATISSE (R Scl and V Hya). The 3.1 $μ$m images presented here show highly asymmetric features, another case of a C-rich star with irregular morphologies close to the stellar disk; this supports the notion that the $C_2H_2+HCN$ abundance distribution usually originates from a clumpy layer around carbon stars.

Figures (11)

• Figure 1: Spectral energy distribution of X TrA. The black points represent the photometric data listed in Table \ref{['phot']}, and the green points correspond to IRAS data. Red and blue data points come from the current MATISSE observations, and the yellow dataset represents the MIDI observations of paladini2017. The black arrows mark some of the spectral features we are interested in.
• Figure 2: AAVSO aavso light curve of X TrA (blue and green symbols), from 5 February 2000 to 13 February 2025. The MIDI and MATISSE epochs of observation are highlighted in orange and red, respectively.
• Figure 3: MiRA image reconstructions for the MATISSE L band for the 3.1 $\mu$m wavelength region (leftmost column), the 3.5 $\mu$m interval (middle), and the reconstructed data for the 3.8 $\mu$m region (right). In the top row, magenta (cyan) contours indicate where the signal is 5 times (30 times) above the estimated background noise, and the inner red contour encapsulates the regions where the S/N is in the top 10% of all values across the image. The blue cross indicates the position of the emission peak seen at 3.8 $\mu$m. In the middle row, the innermost dashed cyan circle represents the photospheric diameter of paladini2017, while the outer circle corresponds to the photospheric diameter scaled to 12 $\mu$m, based on the size ratio of the 2 and 12 $\mu$m mass-losing dynamic models in paladini2009. Both are marked with a dashed blue line in the radial profile plots. The sectors marked by the dotted white lines show the regions used in the plotting of the radial profiles presented in the bottom row. The images were normalised with respect to the highest intensity pixel and re-centred by computing the local average intensity using a square kernel corresponding to 3 mas in radius. The white ellipse shows the estimated beam size and its inclination at each wavelength interval.
• Figure 4: MiRA image reconstructions for the MATISSE N band for the 8.5 $\mu$m (top row) and 11.3 $\mu$m (bottom row) wavelength region. Magenta (cyan) contours indicate where the signal is 5 times (30 times) the estimated background noise, and the inner red contour encapsulates the regions where the S/N is in the top 10% of all values across the image. The innermost dashed cyan circle represents the photospheric diameter from paladini2017, while the outer circle corresponds to the photospheric diameter scaled to 12 $\mu$m, based on the size ratio of the 2 and 12 $\mu$m mass-losing dynamic models in paladini2009. Both are marked with a dashed blue line in the radial profile plots. The sectors marked by the dotted white lines are the regions used in the plotting of the radial profiles presented in the bottom row. The images were normalised with respect to the highest intensity pixel and re-centred on the maximum local average intensity, computed using a square kernel with a radius of 3 mas. The white ellipse shows the estimated beam size and its inclination at each wavelength interval.
• Figure 5: VLTI/MATISSE uv coverage of X TrA, colour-coded according to the telescope configuration used.