FORCASTing the Spectroscopic Dust Properties of the WC+O Binary WR 137 with SOFIA

Massive stars have large impacts on their environments through strong stellar winds and terminal supernova explosions. The exact nature of the impacts can be changed through binary companions and their orbital architecture. These stars are usually formed in binary systems, and a majority of the systems are expected to interact during their lifetimes (Sana et al. 2012, 2013). Among the massive stars, Wolf–Rayet (W-R) stars represent highly evolved massive stars that have lost their outer hydrogen envelope from either strong stellar winds or binary interactions. They are helium burning and have strong hydrogen-free stellar winds ($\dot{M}\sim {10}^{-5}{M}_{\odot }\,{\mathrm{yr}}^{-1};$ v_∞ ≳ 2000 km s⁻¹) and are assumed to have gone through rapid mass loss before becoming a W-R star. These stars have highly structured, strong stellar winds with terminal speeds of several thousand km s⁻¹. From their spectra, the W-R stars are classified into either nitrogen-rich (WN) or carbon-rich (WC) depending on if the wind emission is dominated by nitrogen and ionized helium or carbon.

WR 137 was one of the first W-R stars discovered by Wolf & Rayet (1867) at the Paris Observatory along with WR 134 and WR 135 (where they were labeled by their 1850 coordinates, the present names coming from the "Sixth Catalogue" (van der Hucht et al. 1981). The latter authors classified its spectrum as WC7+abs, reflecting its uncertain status as a binary. From moderate-resolution spectra, Underhill (1962) deduced that it was a WC7+Be shell star binary but Massey et al. (1981) and Moffat et al. (1986) did not find variations in its radial velocities, deducing that WR 137 was not a (close) binary system.

Multiwavelength infrared (IR) photometry of late subtype WC stars found a few to show emission from carbonaceous dust (Allen et al. 1972). Two W-R stars, WR 140 (HD 193793; Williams et al. 1978, WC7+abs) and WR 48a (Danks et al. 1983, WC9) were observed to show rapid increases in the IR attributed to dust formation. IR photometry of WR 137 in 1978–84 (Williams et al. 1985) also showed a dust-formation episode, with maximum near 1984.5. Reexamination of IR photometry of WR 137 in 1970–77 when the emission was fading (Hackwell et al. 1976) suggested that WR 137 was at that time recovering from an earlier, unobserved, dust-formation event, leading Williams et al. (1985) to suggest that such eruptions recurred at intervals of about 15 yr.

The episodic dust formation resembling that by WR 140 revived the search for binary motion. Annuk (1991) demonstrated that the WR 137 system was a binary by presenting a first radial velocity (RV) orbit, adopting a period of 4400 days from the IR data but Underhill (1992) did not accept this and advocated a single star model for WR 137.

Continued IR photometry (Williams et al. 2001) revealed another dust-formation episode peaking in 1997, suggesting a period near 4765 ± 50 days. Lefèvre et al. (2005) derived a spectroscopic period of 4766 ± 66 days from their RV orbit, confirming the binary status of WR 137. Richardson et al. (2016) collected H-band interferometry with the CHARA Array, which allowed them to separate the two components and support a nearly edge-on inclination despite only one epoch of observations. The two stars contribute nearly equally to the H-band flux. They also found a spectral type of WC7pd+O9V using spectroscopic modeling.

The WR 137 dust cloud was imaged in the IR by Marchenko et al. (1999) in 1997 and 1998, shortly after IR maximum, using the Hubble Space Telescope's NICMOS instrument in the H' and K' bands. They found IR-emitting clumps near the central binary, and evidence for a stream of dust emanating from the central source with an extent of ∼025 (∼4 pc). This type of structure was confirmed by Lau et al. (2023) who presented JWST aperture-masking interferometric observations with the NIRISS instrument.

The dust may form in the shocked gas between a WC star and companion O star where their winds collide, such as inferred from the rotating "pinwheel" of heated dust made in the WC9 system WR 104 (which still requires determination of its orbit) and the canonical episodic dust maker WR 140. This system was recently imaged by JWST+MIRI in the mid-IR with three filters spanning 7.7–21 μm, showing more than 150 yr of dust created and sent into interstellar space in concentric arcs (Lau et al. 2022). Spatially resolved spectroscopy of the dust indicated two unidentified IR (UIR) features around 6.4 and 7.8 μm, as well as a strong [S IV] feature at 10.5 μm. This greatly extended the ground-based IR imaging of WR 140's dust by Monnier et al. (2002) and Williams et al. (2009).

A decretion disk around the companion star in WR 137 was discovered from the observation of double-peaked hydrogen and helium emission lines, making it an O9e star. This was first noted by Underhill (1962), and recently investigated in much more detail by St-Louis et al. (2020). This may explain the constant component of the continuum polarization (Harries et al. 2000) and the thinner spiral of dust around WR 137 (St-Louis et al. 2020; Lau et al. 2023).

In this paper, we report on new mid-IR spectroscopy of WR 137 taken with the Stratospheric Observatory for Infrared Astronomy (SOFIA) and the FORCAST instrument as the recent dust formation began. We describe these observations in Section 2. Section 3 describes measurements we made of spectral features in our observations, including a measurement of a line that appears to be associated with the dust being created. We discuss these results with a comparison to other W-R systems and their dust features in Section 4, concluding this study in Section 5.

2.1. Infrared Photometry

2.2. SOFIA Observations

We obtained mid-IR grism spectroscopy using the FORCAST instrument (Herter et al. 2013, 2018) on board SOFIA (Temi et al. 2018). The first observation, taken in 2021 July used both the G063 and G111 grating setups, with two subsequent observations in 2022 February and May ⁹ with only observations using the G063 grism. This setup covers the wavelength range of 4.9–8.0 μm, but we found the spectra near the edge of the chip to be too noisy for analysis, effectively reducing the useful range to 5.1–7.9 μm. These spectra were taken with the 24 slit, yielding a spectral resolving power of ∼180. The FORCAST grism data are known to have variable slit losses, adding uncertainty to the flux calibration, but these data should be accurate to a few percent (e.g., Gehrz et al. 2021). For this analysis, we focus on the variability of line flux and equivalent widths, so we concentrate on only the G063 spectra.

Zoom In Zoom Out Reset image size

Richardson et al. (2016) presented models of the combined spectrum of WR 137 based on optical and ultraviolet spectroscopy. The model, along with the three flux-calibrated spectra from SOFIA are shown in Figure 2. While the SOFIA data are noisy, a few things are readily apparent. The blended lines at 5.4 and 5.9 μm are not present in the models presented by Richardson et al. (2016). The He i line at ∼7.5 μm appears anomalously strong compared with the model or other helium transitions in our spectra and it is possible that hydrogen Pfund-α (7.46 μm) from the decretion disk around the companion star reported by St-Louis et al. (2020) could be a significant contributor. We were able to measure the 6.4 μm feature equivalent width and then compare it to known W-R lines of reasonable strength at 5.6, 5.8, 6.9, and 7.45 μm. Our errors are statistical in nature as we measured the strength with several assumptions of where the continuum was present, allowing us a reasonable confidence on the strengths of the lines measured. In Figure 2, we also show line identifications for the spectrum obtained in 2021 July. We identify lines of He i, He ii, C iii, and C iv, including several that were not included in the PoWR model reported by Richardson et al. (2016).

Zoom In Zoom Out Reset image size

The spectra are compared with the IR light curve in Figure 3, showing that the changing flux observed by SOFIA was seen in the NIR photometry. We show our equivalent width measurements for the 6.4 μm feature as well as other lines in Figure 4. To the first order, we see that the equivalent width of the lines from the wind appear to decrease as the continuum increased through these epochs. This is typical behavior for the wind lines given the additional continuum flux from dust being formed in the wind, as demonstrated in the IR photometry (Figure 1). We also show in Table 1 the net flux for the blended lines at 6.9 μm (the closest line to the UIR band) that shows a constant flux (within uncertainties) across the three observations.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

While the equivalent width of the W-R wind lines decreased during our three observations, the opposite is true for the 6.4 μm feature, which grew stronger with time. This shows that this feature is inconsistent with the W-R wind lines and is therefore associated with the dust production near periastron. We tested this by taking ratios of the 6.4 μm line to each of the wind lines. These ratios, shown in Figure 5, show that this line indeed appears stronger with time, consistent with this feature being formed in the newly formed dust around the system.

Zoom In Zoom Out Reset image size

We have found that a seemingly weak emission line at ∼6.2–6.5 μm (see Figure 6) in the IR spectra of WR 137 taken with SOFIA during its last year of operation. The emission line increased in strength as the dust formation began leading up to the periastron passage that will occur in ∼2024. Chiar et al. (2002) observed a similar feature in the dusty W-R binary WR 48a. In this discussion, we will compare this feature with that seen in other WCd binaries and how its form changed throughout these three observations and what that could imply for dust creation in WC binaries, and in particular for the case of WR 137.

Zoom In Zoom Out Reset image size

Endo et al. (2022) observed a broad 8 μm feature in the spectrum of the 28.1 yr period system WR 125 when the system was near its IR maximum. This feature was seen in the Infrared Space Observatory (ISO) spectra of several WCd stars and was compared to these other systems by Endo et al. (2022). The broad 8 μm feature is often associated with the UIR features, and thus it probably represents emission from the building blocks of the dust particles in the colliding wind region. Our SOFIA spectra of WR 137 are cut off by the detector and do not cover the region of the 8 μm feature.

However, the 8 μm feature is usually associated with a feature at 6.2 μm in IR dust spectra. Duley & Williams (1981) found that the UIR features are often inherent in compounds made of aromatic C−H and C = C chemical bonds, with some suggesting that this arises from polycyclic aromatic hydrocarbons (PAH) emission (e.g., Allamandola et al. 1985). Since the UIR feature is usually seen at 6.2 μm, the observed wavelength of the feature would imply that the feature is shifted to a longer wavelength in the spectrum of WR 137. However, it is also seen at a longer wavelength in the spectrum of WR 48a presented by Chiar et al. (2002), which is classified as "class C" according to the procedures outlined by Peeters et al. (2002).

The longer wavelength of the feature can be caused by hydrogen being absent in the PAH molecules that form this line (e.g., Pauzat et al. 1997). Furthermore, we saw some evidence that the central wavelength of the feature shifted to longer wavelengths during the year of our observations. In Figure 6, we show the spectra in the region of this line with the average peak wavelength shown. For clarity, we also smooth these spectra and shift the last spectrum to a similar flux level as the other observations for clarity. We caution that the signal-to-noise of these spectra is low as evidenced in the unsmoothed spectra in Figure 6, but the overall shift of the line peak went from ∼6.29 μm in 2021 July to ∼6.35 μm in 2022 February, and then to ∼6.41 μm in 2022 May. We speculate that this shift implies that the dust begins forming in the WR 137 system in a mixing environment of the WC wind and the Oe hydrogen disk. As periastron nears, the dust is sent along the collision region which becomes more intense. These initial grains can then grow with additional carbon from the WC wind but with less hydrogen available from the Oe wind.

Lau et al. (2023) presented aperture-masking interferometry of the WR 137 system at 3.8 and 4.8 μm taken with the NIRISS instrument on JWST. The resulting images of the dust around WR 137 show a very narrow extension of the dust compared to the models that are normally able to reproduce the geometry of the WCd binary dust (e.g., WR 112; Lau et al. 2020b). These models have been widely used both for systems with well-established orbits like WR 140 (Lau et al. 2022) as well as for inferring orbital information from long-period systems like Apep (WC8+WN; Callingham et al. 2020), WR 48a (WC6; Lau et al. 2020a), or WR 112 (WC8; Lau et al. 2020b). The narrow signature of dust formation in the WR 137 system could imply that the mixing with the Oe star's disk material is important and crucial for the dust formation in this system, which allows for periodic dust formation as the disk is remarkably stable (St-Louis et al. 2020).

We suggest that the wind of the WC star collides with the equatorial disk surrounding the Oe star. As this begins, the disk provides a large reservoir of hydrogen that mixes with the carbon-rich WC wind. This mixing begins at times where the PAH-like formation can create molecules that are not very hydrogen deficient. However, as the collisions in the system build and the binary approaches periastron, we see the feature move to longer wavelengths indicating a potential for fewer hydrogen atoms to be bound to the PAHs emitting from the feature. As a result of this mixing, the dust is built from the disk, but gradually grows with more carbon-rich material than with the mixed materials.

The IR light curve of WR 137 (Figure 1) has been seen in the past to be somewhat more irregular than other episodic dust-producing WC binaries. In the WR 140 system, the light curve is almost completely repeatable with very few or no excursions. However, in the light curve of WR 137, Williams et al. (2001) show evidence of "minieruptions" happening in 1987, 1988, and 1990 on top of the regular activity. The current outburst shows some deviations from these previous outbursts. These differences from the repeatable behavior of WR 140 shows that variable disk activity from the Oe star could be a protagonist in our understanding of dust production for WR 137.

The observations presented here, especially when coupled to the stellar parameters presented by St-Louis et al. (2020) and the JWST observations presented by Lau et al. (2023) provide some speculative evidence for mixing being an important aspect for dust creation in the WCd binaries with episodic dust creation, or at least for the case of WR 137. Lau et al. (2023) also presented evidence that the collision strength between the two winds for WR 137 was not strong enough for regular dust formation, stating that the disk was required for the dust formation in WR 137 in order to provide an environment with high enough density to begin the dust-formation process.

The disk seems to be an essential ingredient in this systems dust creation, so we also want to consider how such a star evolved in the system. Oe stars, like the Be stars (Rivinius et al. 2013), are rapid rotators. This is also observed in WR 137 where Richardson et al. (2016) measured a $v\sin i$ of 220 km s⁻¹ for the O star, although this is close to an average rotational velocity for a Be star (e.g., Martayan et al. 2007). The results of Marchenko et al. (1999) and Richardson et al. (2016) seem to point toward an edge-on orbital geometry for WR 137, a result that is being confirmed with follow-up interferometry (N. D. Richardson et al. 2023, in preparation), which could mean the Oe star is not aligned with the orbit.

We have found that the mid-IR emission line at 6.3–6.4 μm in the spectrum of WR 137 is associated with the growth of dust during the approach toward periastron passage in 2021–2022. The feature grew in strength as the continuum emission from the dust also grew while the relative strength of the emission lines from the W-R wind appeared to shrink due to the increasing continuum. The peak wavelength of the feature appeared to shift to redder wavelengths as the feature grew, likely associated with hydrogen being less prevalent in the grains as the grains grew. We do caution that this is a marginal detection with the signal-to-noise of the data, but the long dust plume, as recently imaged with aperture-masking interferometry with JWST, is incredibly narrow. We speculate that these results show that the dust grain growth begins in a mixing of material from the WC wind and the Oe disk. Such a mixing is less important as the dust continues to build, leading to a shift in the peak wavelength of the 6.2 μm feature that is typically considered a UIR feature.

SOFIA observed both WR 137 and WR 125 (A. R. Daly et al. 2023, in preparation) during recent dust-producing episodes. These data along with some IR data taken from the ground (e.g., Endo et al. 2022) provide a look into the dust formation at critical epochs in the binary clocks for these systems. Such observations will provide a pivotal context for interpretation of higher-resolution and higher-signal-to-noise spectroscopy of spatially resolved dust around these massive binaries as has already been accomplished with JWST with the prototype of these systems, WR 140 (Lau et al. 2022).

Based on observations made with the NASA/DLR Stratospheric Observatory for Infrared Astronomy (SOFIA). SOFIA was jointly operated by the Universities Space Research Association, Inc. (USRA), under NASA contract NNA17BF53C, and the Deutsches SOFIA Institut (DSI) under DLR contract 50 OK 2002 to the University of Stuttgart. Financial support for this work was provided in part by NASA through award No. 09-0163 issued by USRA. M.J.P. is grateful to Embry-Riddle Aeronautical University's Undergraduate Research Institute for financial support through their IGNITE program. N.D.R. is grateful for support from the Cottrell Scholar Award No. CS-CSA-2023-143 sponsored by the Research Corporation for Science Advancement.

Facility: SOFIA (FORCAST) - Stratospheric Observatory For Infrared Astronomy.

Software: astropy (Astropy Collaboration et al. 2013, 2018).