CASCADE: Filamentary accretion flows in Cygnus X DR20

Astronomy &Astrophysics manuscript no. aa49656-24_appendix ©ESO 2025 November 21, 2025 CASCADE: Filamentary accretion flows in Cygnus X DR20 M. Sawczuck1, H. Beuther1, S. Suri1,2, F. Wyrowski3, K.M. Menten3?, J.M. Winters4, L. Bouscasse4, N. Schneider5, T. Csengeri6, C. Gieser1,7, S. Li1, D. Semenov1, I. Skretas3, M.R.A. Wells1, 1Max Planck Institute for Astronomy, Königstuhl 17, 69117 Heidelberg, Germany, email: [email protected] 2Department of Astrophysics, University of Vienna, Türkenschanzstrasse 17,1180 Vienna, Austria 3Max-Planck-Institut für Radioastronomie, Auf dem Hügel 69, 53121 Bonn, Germany 4IRAM, 300 rue de la Piscine, Domaine Universitaire de Grenoble, 38406 St.-Martin-d’Hères, France 5I. Physik. Institut, University of Cologne, Cologne, Germany 6Laboratoire d’astrophysique de Bordeaux, Univ. Bordeaux, CNRS, B18N, allée Geoffroy Saint-Hilaire, 33615 Pessac,France 7Max Planck Institute for Extraterrestrial Physics, Gießenbachstraße 1, 85749 Garching bei München, Germany ABSTRACT Context. Filamentary gas flows are an important process to funnel gas from cloud scales onto star-forming cores. Aims. We investigate the role of filaments in high-mass star formation, whether gas flows from large to small scales along them, and what their properties might reveal about the region they are found in. Methods. The Max Planck IRAM Observatory Program (MIOP): The Cygnus Allscale Survey of Chemistry and Dynamical Environments (CASCADE) includes high spatial resolution (∼300) data of HCO+(1 −0) and H13CO+(1 −0) emission in the star-forming DR20 region in the Cygnus X complex. In this data we identify filaments with the structure identification algorithm DisPerSE. We further analyse these filaments using Gaussian fits to the spectra to determine the line peak velocity and full-width-half-maximum along them. The Python package FilChaP is used to determine filament widths. Results. We find projected velocity gradients inside several filaments between 0.4 to 2.4 km s−1over projected length-scales of 0.1 pc towards star-forming cores. This can be interpreted as a sign of gas flowing along the filaments towards the cores. The filament width distributions exhibit median values between 0.06 and 0.14 pc depending on core, tracer and method. Standard deviations are approximately around 0.02 to 0.06 pc. These values are roughly in agreement with the filament width of 0.1 pc typically found in nearby low-mass star-forming regions. Conclusions. This first analysis of filamentary properties within the Cygnus X CASCADE program reveals potential signatures of gas flows along filaments onto star-forming cores. Furthermore, the characteristics of the filaments in this high-mass star-forming region can be compared to those of filaments in low-mass star-forming regions typically studied before. Extending such studies to the entire CASCADE survey will enhance our knowledge of high-mass filament properties on solid statistical grounds. This contribution provides additional supplementary material to the paper to be published in Astronomy and Astrophysics, 2026. Key words. stars: formation – ISM: structure – ISM: kinematics and dynamics – ISM: individual objects: Cygnus X – ISM: individual objects: DR20 ?In memory of Karl Menten, who suddenly passed away before completing this work. His invaluable advice and contributions will be deeply missed. Article number, page 1 of 38 A&A proofs: manuscript no. aa49656-24_appendix Appendix A: Velocity and linewidth of each region Appendix A.1: HCO+ Core A: For core A in HCO+(s. 1st row in Fig. 8 of the main paper, Sawczuck et al. 2026, and Figs. B.1,B.2), 4 out of 5 filaments show a velocity gradient close to the core, with a change in velocity of 0.8 to 1.3 km s−1within 0.1 pc distance. Note, however, that the filaments come together at the core position, and their first spectra at 0 pc distance are often the same for several filaments. Here, this is the case for A3 to A5. Nevertheless, two filaments connected to core A independently show a velocity gradient (A1 and A3). In contrast, the velocity of A2 is not linear close to the core. While not enough third peaks are identified in A5 to show them in separate plots, it is a good example of a filament with possibly three velocity components. While over the length of each filament of core A the linewidth shows large variance, it is approximately 2 km s−1 close to the core for each velocity component of each filament, with the one exception of one high velocity component (s. A1 in Fig. B.1), where it is ∼3 km s−1. The mean linewidth of all filaments of core A is between 1.7 and 2.2 km s−1, with standard deviations between 0.3 and 0.7 km s−1. Core B: Core B in HCO+(s. 2nd row in Fig. 8 of the main paper, Sawczuck et al. 2026, (B3) and Fig. B.3) is the only core in HCO+where no filament shows a velocity gradient within 0.1 pc distance to it. Instead, B1 and B3 (s. 2nd row Fig. 8 of the main paper, Sawczuck et al. 2026, and Fig. B.3) show an almost constant velocity over their whole distance for their main velocity component. B2 shows the absolute velocity at 0.1 pc first become higher towards the core, but then become lower again (i.e. non-linear velocity). We can also see this trend in other cores (e.g., D3 and D4 in HCO+, Fig. B.5). Apart from the first 0.1 pc close to the core, in contrast to B1 and B3, the velocity of B2 rises gradually over its whole 1 pc length. This is also a trend we see in several cores in HCO+(e.g., D2 to D4, Fig. B.5), although most other filaments are shorter. The linewidth of the filaments of this core have a high variance over the filament length very similar to core A. Close to the core it is approximately 2 km s−1for two of the four filaments (B2 and B4), 3 km s−1for one (B1) and close to 4 km s−1for B3. The mean linewidth of all filaments of core B is between 1.6 and 2.5 km s−1, with standard deviations between 0.5 to 0.7 km s−1. Core C: C1 shows a constant velocity over its whole short length (∼0.1 pc). The velocity of C2 is gradually rising over the filament length, but shows a negative velocity gradient close to the core (i.e., the absolute velocity becomes lower towards the core). This negative gradient is the only one we find, and seems to be connected to a third velocity component. The velocity of C3 rises at first with a gradient of 0.5 km s−1over 0.1 pc, keeps rising for another ∼0.1 pc and then stays approximately constant. The velocity of C4 in HCO+(Fig. B.4) shows a gradient of ∼0.8 km s−1over 0.1 pc of the core, and keeps gradually rising over the whole length of the filament. The linewidth of the main velocity components is approximately 2 km s−1at the core for every filament. The mean linewidth of all filaments of core C is between 1.6 and 2.4 km s−1, with standard deviations between 0.5 and 0.7 km s−1, which is very similar to core B. Core D: D1 (and D2 with the same spectra at the core) in HCO+ (Fig. B.5) shows a velocity gradient of ∼0.4 km s−1over 0.1 pc. D3 and D4 show non-linear velocities close to the core. Three out of the four filaments (D2 to D4) have a gradually rising velocity from 0.1 pc to larger distances. The linewidth of D4 shows a large scatter, with a mean of 2.2 km s−1and a standard deviation of 0.9 km s−1, while D2 shows an almost constant linewidth up to a distance of ∼0.3 pc. The linewidth at the core of the main velocity component is below 2 km s−1for all filaments. The mean linewidth of all filaments of core D is between 1.8 and 2.2 km s−1, with a standard deviation between 0.6 and 0.9 km s−1. The filament origins together with the intensity peak of HCO+are offset from the continuum source in this region by a distance on the order of 10−2pc. Core E: Core E in HCO+(Fig. B.6) shows a large velocity gradient in all filaments of 2.0 km s−1(E2-E4 with the same spectra) to 2.4 km s−1(E1), possibly originating from two velocity components starting at ∼-3.5 and ∼-1.5 km s−1at the core, respectively. The velocity of E3 keeps rising over the whole filament length. The linewidth at the core is ∼2 km s−1for all filaments and then rises (possibly due to merging components). Over the whole filament length it shows a large scatter with mean values between 1.9 and 2.4 km s−1and standard deviations between 0.7 and 0.9 km s−1. The filament origins in this region are even further offset from the continuum source than in core D. Core F: Due to its low intensity in HCO+, there are many points missing in the velocity and linewidth of core F, especially close to the core. Therefore, we refrain from analysing the results (Fig. B.7). Remote A: The velocity and linewidth of filament 1 in the remote area around core A (Fig. B.8) have little variation. The mean linewidth is 1.7 km s−1with a standard deviation of 0.4 km s−1. Arem2 (Fig. 8 of the main paper, Sawczuck et al. 2026) shows three velocity components and larger scatter of each as well as of the linewidth. The mean linewidth and its scatter of the last two filaments is even higher with 2.3 km s−1and a standard deviation of 0.8 km s−1. Remote B: The filament close but not connected to core B in HCO+(Fig. B.9) is the only filament that shows a clear third peak in more than 10 spectra along the filament. However, it is difficult to separate the peaks into distinct components, and it is unclear if there are three velocity components or more. A large range of velocities is observed in this filament, as well as a large range of linewidths, with a mean of 1.9 km s−1and a standard deviation of 0.8 km s−1. Appendix A.2: H13CO+ Core A: Of the 6 filaments connected to core A in H13CO+ (Fig. 9 of the main paper, Sawczuck et al. 2026, and Figs. B.10, B.11), close to the core most of them share the same spectra (A1 with A3, and A2 with A4-A6). A1 and A3 show a velocity gradient of 1.7 km s−1. As in core E in HCO+, there may be two velocity components close to the core fitted as one, resulting Article number, page 2 of 38 M. Sawczuck et al.: CASCADE: Filamentary accretion flows in Cygnus X DR20 in a large linewidth and a velocity gradient as the higher velocity component becomes more prominent. For the other filaments (A2, A4-A6) the velocity first becomes lower and then higher from the core outward within 0.1 pc (i.e. non-linear velocity). A2 shows an approximately constant velocity over its whole length. All filaments in core A show a linewidth of 2.5 to 3.5 km s−1 at the core, which then becomes lower (∼1 km s−1) further away from the core. The mean linewidth of all filaments is between 1.4 and 1.8 km s−1, with standard deviations between 0.5 and 0.7 km s−1. Core B: In core B in H13CO+(Fig. B.12) both filaments share the first few spectra close to the core, which show a velocity gradient of 1.4 km s−1over 0.1 pc. This gradient together with the rising linewidth close to the core could be a sign that similar to core E in HCO+and core A in H13CO+, there are two velocity components merging together. However, with visual inspection we cannot confirm this. The linewidth rises from 1.9 km s−1at the core up to 3.3 km s−1over 0.04 pc distance from the core outward and then decreases to approximately 1 km s−1further from the core. The mean linewidth of the two filaments is 1.6 and 1.8 km s−1, with standard deviations of 0.7 and 0.8 km s−1. Core C: C1 in H13CO+(Fig. B.13) shows a velocity gradient close to the core of 0.7 km s−1over 0.1 pc. The other two filaments share the spectra closest to the peak, which show no velocity gradient within 0.1 pc from the core. Although the velocity close to the core of C2 is non-linear, it rises over its whole length. The linewidths of all filaments are ∼2 km s−1at the core and then become lower but with large scatter. The mean linewidth of all filaments is between 1.3 and 2.1 km s−1, with standard deviations between 0.4 and 0.6 km s−1. Core D: Three of the filaments of core D (D1, D2, D4) in H13CO+(Fig. B.14), where two of them share the same spectra close to the core (D1 and D2), show velocity gradients of ∼0.7 to 0.8 km s−1over 0.1 pc. D3 shows an approximately constant velocity over its whole (short) length. The linewidth is rather low and constant between around 1 and 2 km s−1, with the exception of D4, which shows a large scatter further from the core. The mean linewidth of all filaments is between 1.2 and 1.6 km s−1, with standard deviations between 0.2 and 0.8 km s−1. Core E: E1 in H13CO+(Fig. B.15) shows a slight overall increase of the velocity along the filament. Close to the core its velocity rises and falls within 0.1 pc, and the other way around for E2, i.e. both filaments show non-linear velocities. The linewidth of both filaments decreases from ∼2 km s−1 at the core to ∼1 km s−1further away, with larger values in between and large errors, with means of 1.6 and 1.4 km s−1and a standard deviation of 0.7 and 0.4 km s−1. Core F: Core F in H13CO+(Fig. B.16) has too few points and large errors to analyse the results. Appendix B: Additional figures Additional figures shown are: –Fig. B.1 to Fig. B.16: Linewidth and velocity within all analysed filaments (s. Section 4.1 in Sawczuck et al. 2026). –Fig. B.17 to Fig. B.29: Histograms of filament width distributions (FilChaP results) for all methods and all cores (s. Section 4.2 in Sawczuck et al. 2026). –Fig. B.30 to Fig. B.45: Filament widths (FilChaP results) along all filaments for all methods (s. Section 4.2 in Sawczuck et al. 2026). Article number, page 3 of 38 A&A proofs: manuscript no. aa49656-24_appendix Fig. B.1: Rows: Filaments A1 to A2 (core A). Images: HCO+(1 −0) 0th moment map. The white dot within the core shows the coordinates of the continuum source, while the yellow dot shows the filament origin. The filament corresponding to the plots is shown in its velocity colours. Plots: Gaussian fitted FWHM (left) and peak positions (right) of the spectra within the corresponding filament, plotted over the distance from the core. The yellow dot corresponds to a distance of zero. If there is a significant amount of data points (≥10) for a second spectral Gaussian peak, the positions of the higher velocity peaks (top) and the lower velocity peaks (bottom) are shown in separate plots. The peaks that do not clearly belong to either of the two components are shown in gray in the closer one. The other data points are coloured by their fitted peak positions. All y-ranges of the same type of plots of the same core are the same for better comparison between the filaments (note that they are different for each core). Article number, page 4 of 38 M. Sawczuck et al.: CASCADE: Filamentary accretion flows in Cygnus X DR20 Fig. B.2: Rows: Filaments 3-5 of core A in HCO+(1 −0). For more details s. preceding figure (Fig. B.1). Article number, page 5 of 38 A&A proofs: manuscript no. aa49656-24_appendix Fig. B.3: Rows: Filaments 1-4 of core B in HCO+(1 −0). For more details s. Fig. B.1. Article number, page 6 of 38 M. Sawczuck et al.: CASCADE: Filamentary accretion flows in Cygnus X DR20 Fig. B.4: Rows: Filaments 1-4 of core C in HCO+(1 −0). For more details s. Fig. B.1. Article number, page 7 of 38 A&A proofs: manuscript no. aa49656-24_appendix Fig. B.5: Rows: Filaments 1-4 of core D in HCO+(1 −0). For more details s. Fig. B.1. Article number, page 8 of 38 M. Sawczuck et al.: CASCADE: Filamentary accretion flows in Cygnus X DR20 Fig. B.6: Rows: Filaments 1-4 of core E in HCO+(1 −0). For more details s. Fig. B.1. Article number, page 9 of 38 A&A proofs: manuscript no. aa49656-24_appendix Fig. B.14: Rows: Filaments 1-4 of core D in H13CO+(1 −0). For more details s. Fig. B.1. Article number, page 16 of 38 M. Sawczuck et al.: CASCADE: Filamentary accretion flows in Cygnus X DR20 Fig. B.15: Rows: Filaments 1-2 of core E in H13CO+(1 −0). For more details s. Fig. B.1. Fig. B.16: Rows: One filament close to but not directly connected to core F in H13CO+(1 −0). For more details s. Fig. B.1. Article number, page 17 of 38 A&A proofs: manuscript no. aa49656-24_appendix Fig. B.17: Deconvolved widths of slices of filaments connected to core A in HCO+determined with FilChaP shown in four histograms for the four different methods used. Mean, median and standard deviation for each distribution is given in each plot. The filaments used for this calculation are coloured in the image by the velocity of the channel they were identified in. Fig. B.18: Deconvolved widths of slices of filaments connected to core B in HCO+determined with FilChaP. For further details s. Fig. B.17. Article number, page 18 of 38 M. Sawczuck et al.: CASCADE: Filamentary accretion flows in Cygnus X DR20 Fig. B.19: Deconvolved widths of slices of filaments connected to core C in HCO+determined with FilChaP. For further details s. Fig. B.17. Fig. B.20: Deconvolved widths of slices of filaments connected to core D in HCO+determined with FilChaP. For further details s. Fig. B.17. Article number, page 19 of 38 A&A proofs: manuscript no. aa49656-24_appendix Fig. B.21: Deconvolved widths of slices of filaments connected to core E in HCO+determined with FilChaP. For further details s. Fig. B.17. Fig. B.22: Deconvolved widths of slices of filaments connected to core F in HCO+determined with FilChaP. For further details s. Fig. B.17. Article number, page 20 of 38 M. Sawczuck et al.: CASCADE: Filamentary accretion flows in Cygnus X DR20 Fig. B.23: Deconvolved widths of slices of filaments in the remote area close to core A in HCO+determined with FilChaP. For further details s. Fig. B.17. Fig. B.24: Deconvolved widths of slices of the filament close but not connected to core B in HCO+determined with FilChaP. For further details s. Fig. B.17. Article number, page 21 of 38 A&A proofs: manuscript no. aa49656-24_appendix Fig. B.25: Deconvolved widths of slices of filaments connected to core A in H13CO+determined with FilChaP. For further details s. Fig. B.17. Fig. B.26: Deconvolved widths of slices of filaments connected to core B in H13CO+determined with FilChaP. For further details s. Fig. B.17. Article number, page 22 of 38 M. Sawczuck et al.: CASCADE: Filamentary accretion flows in Cygnus X DR20 Fig. B.27: Deconvolved widths of slices of filaments connected to core C in H13CO+determined with FilChaP. For further details s. Fig. B.17. Fig. B.28: Deconvolved widths of slices of filaments connected to core D in H13CO+determined with FilChaP. For further details s. Fig. B.17. Article number, page 23 of 38 A&A proofs: manuscript no. aa49656-24_appendix Fig. B.29: Deconvolved widths of slices of filaments connected to core F in H13CO+determined with FilChaP. For further details s. Fig. B.17. Article number, page 24 of 38 M. Sawczuck et al.: CASCADE: Filamentary accretion flows in Cygnus X DR20 Fig. B.30: Deconvolved widths of filament slices (filaments 1-2) of core A in HCO+determined with FilChaP. Rows: Different filaments leading to the central core. Images: Core A. The white dot within the core shows the coordinates of the continuum source, the yellow dot shows the filament "origin", corresponding to a distance of zero. The filament corresponding to the plot is coloured by the velocity of the channel it was identified in. The green colour shows which other filaments in this region have been analysed; the ones not shown here can be found in the Appendix (Figs. B.30 to B.45). Plot: The deconvolved width of the filament slices is plotted over the distance from the core (or more precisely, the filament "origin", yellow dot in the image). A distance of zero thereby corresponds to the slice closest to the core. The errorbars show the deconvolved FWHM errors resulting from the errors of the corresponding parameters of the respective fit to the radial profile given in the covariance matrix of the curve_fit function used in FilChaP. Article number, page 25 of 38 A&A proofs: manuscript no. aa49656-24_appendix Fig. B.37: Deconvolved widths of filament slices of the remote region close to core A in HCO+determined with FilChaP. A distance of zero in the plots corresponds to the position of the yellow dot. For further details s. Fig. B.30. Article number, page 32 of 38 M. Sawczuck et al.: CASCADE: Filamentary accretion flows in Cygnus X DR20 Fig. B.38: Deconvolved widths of filament slices of the remote region close to core B in HCO+determined with FilChaP. A distance of zero in the plots corresponds to the position of the yellow dot. For further details s. Fig. B.30. Fig. B.39: Deconvolved widths of filament slices (filaments 1-3) of core A in H13CO+determined with FilChaP. For further details s. Fig. B.30. Article number, page 33 of 38 A&A proofs: manuscript no. aa49656-24_appendix Fig. B.40: Deconvolved widths of filament slices (filaments 4-5) of core A in H13CO+determined with FilChaP. For further details s. Fig. B.30. Article number, page 34 of 38 M. Sawczuck et al.: CASCADE: Filamentary accretion flows in Cygnus X DR20 Fig. B.41: Deconvolved widths of filament slices of core B in H13CO+determined with FilChaP. For further details s. Fig. B.30. Article number, page 35 of 38 A&A proofs: manuscript no. aa49656-24_appendix Fig. B.42: Deconvolved widths of filament slices of core C in H13CO+determined with FilChaP. For further details s. Fig. B.30. Article number, page 36 of 38 M. Sawczuck et al.: CASCADE: Filamentary accretion flows in Cygnus X DR20 Fig. B.43: Deconvolved widths of filament slices of core D in H13CO+determined with FilChaP. For further details s. Fig. B.30. Article number, page 37 of 38 A&A proofs: manuscript no. aa49656-24_appendix Fig. B.44: Deconvolved widths of filament slices of core E in H13CO+determined with FilChaP. For further details s. Fig. B.30. Fig. B.45: Deconvolved widths of filament slices of core F in H13CO+determined with FilChaP. For further details s. Fig. B.30. Article number, page 38 of 38