Online material

Appendix A: Individual epochs and summary of radial velocity/variability analysis

Appendix B: Notes on individual ARGUS sources

	Fig. B.1 Non-variable ARGUS sources displaying composite spectra. Emission lines labelled are N iv λ4058 and Si iv λλ4089, 4116. Absorption lines labelled are He i λ4026, He i λ4143, He ii λ4200, He i λ4388, He i λ4471, He ii λ4542 and a diffuse interstellar band at 4428 Å.
Open with DEXTER

We comment here on selected individual sources, paying particular attention to sources with previous identifications, composite spectra, and also to those which appear multiple in the WFC3 image.

• VFTS 542: this star is identified as a definite variable with a largeamplitude from both ARGUS and Medusa observations. ItsHe ii λ4200 and He ii λ4542 lines show a weak P Cygni component and it is classified as O2 If*/WN5 (Paper I), so even if it was not variable, its absolute RV could not be trusted. RV discrepancies as large as ~40 km s^-1 are found at some epochs between He ii λ4200 and He ii λ4542.

• VFTS 545: it is also known as Mk35 (Melnick 1985) and classified as O2 If*/WN5 (Paper I). There is a discrepancy of ~20 km s^-1 between He ii λ4200 and He ii λ4542, and its absolute RV also cannot be trusted. Low-amplitude variability is only detected in He ii λ4542, which is stronger and has smaller RV uncertainties compared to He ii λ4200.

• VFTS 570: the ARGUS spectra of this source were not analysed for RV variability because their S/N was too low, but from the Medusa spectra it was found to be a definite RV variable with a large amplitude (Paper VIII). Two stars appear to contribute significantly to the ARGUS source when comparing with the WFC3 image.

• VFTS 585: this source was also found to be variable from the Medusa spectra (Paper VIII). Significant RV variability was detected from the relatively low S/N ARGUS spectra only once He ii λ4200 and He ii λ4542 were fitted simultaneously.

• VFTS 1001: this source corresponds to a known Wolf-Rayet star, R134 (Feast et al. 1960), classified as WN6(h) (e.g. Massey & Hunter 1998). Although it was detected as an X-ray source and suggested as a possible colliding-wind binary by Portegies Zwart et al. (2002), it is not known to be a binary. Interestingly, our TVS analysis reveals significant variability in the He ii λ4542 emission, but it is unclear if this is due to a normalisation problem in a spectral range dominated by several emission lines, where the continuum is harder to define.

• VFTS 1003: this source was found to be a new B[e]-type star in Paper I. The TVS analysis performed in the present work did not reveal any significant variability other than in the nebular emission lines (due to sky subtraction).

• VFTS 1004: the WFC3 image suggests that two sources are contributing to VFTS 1004, but it does not display a composite spectrum, it is not found to be variable, and the He i λ4388, He ii λ4200 and He ii λ4542 lines all have consistent absolute RVs.

• VFTS 1007: similarly to VFTS 1004, VFTS 1007 appears multiple when inspecting the WFC3 image, but it is not variable, it does not have a composite spectrum, and He i λ4388, He ii λ4200 and He ii λ4542 all have consistent absolute RVs.

• VFTS 1014: the presence of N iv λ4058 and Si iv λ4089/4116 emission together with weak but well developed He i singlet lines at 4121, 4143 and 4388 Å suggests a composite spectrum (see Fig. B.1). Based on the helium line diagnostics and the absence of Si iii λ4552, the later component is identified as a mid/late O-type star. From the relative strength of N iv and Si iv, the other component is O2-4.5 (we cannot be more precise because our the ARGUS spectrum does not include the N v absorption region), in agreement with the O3 V classification of Massey & Hunter (1998). Even though its spectrum appears composite, this source did not show significant variability. The He i λ4388, He ii λ4200 and He ii λ4542 lines all have consistent absolute RVs.

• VFTS 1015: this source is clearly multiple by comparison with the WFC3 image and significant RV variability is found in He ii λ4542. For some epochs, the RV of the He i λ4388 line is clearly different from that of He ii λ4200 and He ii λ4542 lines.

• VFTS 1017: this source is variable. Its He ii λ4200 and He ii λ4542 lines have a weak P Cygni component. A discrepancy of up to ~30 km s^-1 is found in the RVs of He ii λ4200 and He ii λ4542 at different epochs.

• VFTS 1018: the presence of weak N iv λ4058 and Si iv λ4116 emission in combination with weak but well developed He i singlet lines suggests a composite spectrum (see Fig. B.1). Based on the helium line diagnostics and the absence of Si iii λ4552, the later component is identified as a mid/late O-type star. From the relative strength of the N iv and Si iv emission, the other component is classified as O2-4.5, in relatively good agreement with the O3 III(f*) classification of Massey & Hunter (1998). Even though its spectrum appears composite, this source did not show significant variability. The He i λ4388, He ii λ4200 and He ii λ4542 lines all have consistent absolute RVs.

• VFTS 1019: this is a known high-mass binary (R136-038) classified as O3 III(f*) + O8 by Massey & Hunter (1998), then revised as O3 V + O 6 V by Massey et al. (2002). The ARGUS spectra show obvious variability, a large RV amplitude, and SB2 profiles at several epochs.

• VFTS 1022: this source corresponds to Mk37a = R136-014 (Melnick 1985; Massey & Hunter 1998), classified as O4 If+ by Massey & Hunter (1998), but suggested as O3.5 If*/WN7 by Crowther & Walborn (2011). 13 epochs (the source is on the edge of the A3 and A4 ARGUS pointings) made it possible to detect low-amplitude RV variability in He ii λ4542. However, even if it had not been flagged as variable, this star would not have been suitable for our analysis of the dynamics. A discrepancy of ~15 km s^-1 is found between the RVs of He ii λ4200 and He ii λ4542, and the RV of He i λ4388 is significantly larger than that of the He ii lines.

• VFTS 1023: we classified this star as O8 III/V. Massey & Hunter (1998) classified it as O6, but at this subtype He i+ii λ4026 should be as deep as He ii λ4200 while He i λ4471 should be significantly weaker than He ii λ4542. Also, He i λ4143 and He i λ4388 should be much weaker than He ii λ4200 and He ii λ4542 respectively, which is not what we see. A possible explanation for the discrepancy between our classification and that of Massey & Hunter (1998) is that this source is an undetected single-lined spectroscopic binary.

• VFTS 1025: this source appears multiple and the centre of the ARGUS position is offset between two stars in the WFC3 image, with the much brighter star being R136c. It is interesting to note that we find significant variability in the TVS of this source (see Fig. 3). R136c was identified as a probable binary (Schnurr et al. 2009) and suspected to be a colliding-wind massive binary (Crowther et al. 2010).

• VFTS 1026: when comparing with the WFC3 image, the centre of the ARGUS source appears offset between two stars. One of these is MH41, O3 III(f*) (Massey & Hunter 1998), also classified as O8: V by Walborn & Blades (1997). The light is probably dominated by MH41 (the brighter of the two stars), although we flagged VFTS 1026 as having a composite spectrum (see Fig. B.1), as suggested by the presence of N iv λ4058 and Si iv λ4089/4116 emission together with weak but well developed He i singlet lines at 4121, 4143 and 4388 Å. Based on the helium line diagnostics and the absence of Si iii λ4552, the later component is identified as a mid/late O-type star. From the relative strength of the N iv and Si iv emission, the other component is classified as O2-4.5, in agreement with the classification of Massey & Hunter (1998). This source is however not variable, and its He i λ4388, He ii λ4200 and He ii λ4542 lines have consistent absolute RVs.

• VFTS 1031: this corresponds to R136-025 (O3 V; Massey & Hunter 1998), which was flagged as a suspected variable by Massey et al. (2002). In our ARGUS spectra, the He i+He ii λ4026, He ii λ4200 and He ii λ4542 lines seem to display three components at some epochs.

• VFTS 1034: this corresponds to Mk32, which is itself a blend of R136-013 (O8 III(f), Massey & Hunter 1998; O7.5 II, Walborn & Blades 1997) and R136-074 (O6 V, Massey & Hunter 1998). The variability in this source is more obvious in the He i+He ii λ4026 and He i λ4471 lines, but also significant when He i λ4388, He ii λ4200 and He ii λ4542 are fitted simultaneously.

© ESO, 2012