Search for particle acceleration in two massive Wolf-Rayet stars using uGMRT observations

Studies of massive stars are of prime importance in stellar astrophysics given their strong mechanical, radiative, and chemical feedback. In particular, their intense radiation field expels and accelerates the outer layers of the star, giving rise to powerful supersonic winds. These winds deposit a large amount of mechanical energy into the interstellar medium (ISM) generating strong adiabatic shocks suitable for relativistic particle acceleration, presumably via diffusive shock acceleration (DSA; e.g. Drury, 1983, and references therein). Detection of non-thermal (NT) radio emission, which is synchrotron radiation produced by relativistic electrons in the presence of a magnetic field (Blumenthal & Gould, 1970; White, 1985), is an observational evidence of particle acceleration.

About 50 NT radio-emitting binary systems, referred to as particle-accelerating colliding-wind binaries (PACWBs), have been identified¹¹1Updated PACWB catalogue can be found at https://www.astro.uliege.be/~debecker/pacwb/ (De Becker & Raucq, 2013). It is not clear how efficient these systems are at converting wind kinetic energy into particle acceleration, but at least for the WR binary Apep, it was recently shown that this fraction can be close to 1 percent (del Palacio et al., 2023), while for WR 146 it can be as high as 30 percent (Pittard et al., 2021). In an observational breakthrough, NT radio emission from the stellar bubble G2.4+1.4, associated with the presumably single WR star WR 102, was detected by Prajapati et al. (2019). In this case, the fraction of the kinetic wind power that is converted into cosmic-ray acceleration was estimated to be of the order of a few percent. Further, NT radio emission has also been detected in bow-shocks of massive runaways stars (Benaglia et al., 2010; Moutzouri et al., 2022). For the runaway star BD+43 3654, Benaglia et al. (2021) estimated that $\sim 10$ percent of the wind kinetic power should be converted into relativistic particle acceleration in the bow shock. These observations support the hypothesis that massive stars are relevant sources of Galactic cosmic rays, as estimated by Seo et al. (2018) under the assumption that typically 1–10 percent of their wind luminosity is transferred to relativistic particles. Moreover, WR stars can also be key to explain the composition of cosmic rays (Gabici, 2023, and references therein).

The energy budget for NT particle acceleration depends on the wind kinetic power, $P_{\mathrm{kin}}$. For a star with mass-loss rate $\dot{M}$ and terminal velocity $v_{\infty}$, it is $P_{\mathrm{kin}}\approx 0.5\,\dot{M}\,v_{\infty}^{2}$. In addition, under certain assumptions, the efficiency of particle acceleration scales with the square of the shock (and therefore the wind) velocity (Drury, 1983). WR stars, which are the evolved counterparts of O-type stars, have high mass-loss rates $\sim(1$–$5)\times 10^{-5}\,\rm M_{\odot}\,yr^{-1}$ (e.g., Abbott et al., 1986; Leitherer et al., 1997; Chapman et al., 1999) and fast winds with velocities ranging between $\sim 700$–$6000\,\rm km\,s^{-1}$ (e.g., Nugis et al., 1998; Hamann et al., 2000; Nugis & Lamers, 2000). Thus, their winds are exceptionally powerful with a typical $P_{\mathrm{kin}}>10^{37}\,\rm erg\,s^{-1}$. This amounts to substantial mechanical energy available to drive particle acceleration, thus making this subset of hot and massive stars ideal targets to search for particle acceleration.

The requirement of large $P_{\mathrm{kin}}$ sets the primary target selection criteria for this radio investigation of particle acceleration in massive WR stars. Furthermore, there is growing observational evidence of a strong correlation between the detection of NT emission and binarity in O-type and WR systems (e.g. De Becker & Raucq, 2013). In comparison, detection of NT emission in low radio frequency is reported for only one presumably single non-runaway star system (WR 102; Prajapati et al., 2019). This detection opens a new window to probe shock physics and assess the role of the local ISM and its interaction with the powerful WR winds as a necessary ingredient for particle acceleration in single massive stars.

With the aim to explore the above two scenarios, we carried out low-frequency (735 and 1260 MHz) radio observations of WR 114 and WR 142, two nearby (distance $\rm\lesssim 2~{}kpc$) WR stars with large wind kinetic power ($\rm\sim 10^{38}\,erg\,s^{-1}$). Following the discussion in Benaglia et al. (2021), we derived the tangential velocities of these two stars with respect to its local ISM from the latest proper motion values listed in Gaia DR3 database (see Table 1). The estimated tangential velocities are 10.2 and 18.0 $\rm km\,s^{-1}$ for WR 114 and WR 142, respectively, implying that these are not runaway stars (e.g. Eldridge et al., 2011). The coordinates and relevant parameters of these stars are listed in Table 1. Located in the Ser OB1 association, WR 114 is classified as WC5 spectral type. Comparing the strength of emission lines with those in other stars of the same spectral type, van der Hucht (2001) proposed the presence of an OB companion. However, the Potsdam WR spectral modelling was consistent with a single star (Sander et al., 2012). WR 114 was observed with XMM-Newton (Oskinova et al., 2003). The non-detection of X-ray emission ($L_{\rm X}\lesssim 2.5\times 10^{30}\rm erg\,s^{-1}$) was attributed to the high opacity of the metal-rich and dense wind from the WC star. The other target, WR 142 (spectral type WO2), belongs to the Berkeley 87 cluster and is one of the only four WO stars detected in our Galaxy. XMM-Newton and Chandra observations reveal weak X-ray emission with excess absorption below 2 keV (Sokal et al., 2010). These authors discussed several possible emission processes, including thermal and non-thermal ones (e.g., inverse Compton scattering) and a possible colliding-wind shock scenario, to interpret their observations, but their results were not conclusive.

This paper is organized as follows. The observations and data reduction procedure are described in Sect. 2. After reporting on the main results in Sect. 3, we detail in Sect. 4 the scientific discussion of the two WR stars in the framework of the scenarios introduced above. Our main conclusions are finally provided in Sect. 5.

We probed the radio continuum emission associated with WR 114 and WR 142 using the uGMRT situated in Pune, India. GMRT offers a Y-shaped array configuration of 30 fully steerable antennas of 45-m diameter. The central square km houses 12 randomly placed antennas and the remaining 18 antennas are uniformly distributed in the arms, with 6 on each arm. With this hybrid array arrangement, GMRT provides largest and smallest baselines of 25 km and 100 m, respectively, which makes it capable of probing radio emission at both small and large spatial scales. The largest detectable structures in Band 4 and 5 are 17 and 7 arcmin, respectively²²2GMRT Observer’s Manual (http://www.ncra.tifr.res.in/ncra/gmrt/gmrt-users/observing-help/manual_7jul15.pdf). Details of the GMRT system can be found in Swarup et al. (1991) and Gupta et al. (2017).

Dedicated observations (Project code: 38_070) were carried out for both sources in Band 4 (550–950 MHz) and Band 5 (1050–1450 MHz) with the GMRT Wideband Backend (GWB) correlator configured to have a bandwidth of 400 MHz across 4096 channels. Primary calibrators were observed at the beginning and end of the observation for flux and bandpass calibration. The phase calibrators were observed after each scan (30 mins) of the target to calibrate the phase and amplitude variations over the entire observing period. Details of the observations are given in Table 2.

The data reduction process of flagging, calibration, imaging, and self-calibration was done using the CAPTURE³³3https://github.com/ruta-k/CAPTURE-CASA6.git continuum imaging pipeline for uGMRT (Kale & Ishwara-Chandra, 2021), which utilizes tasks from Common Astronomy Software Applications (CASA, McMullin et al., 2007). The Perley & Butler (2017) scale was employed to set the flux density calibration. After the initial rounds of editing and calibration, we used the multi-term multi-frequency synthesis (MT-MFS; see Rau & Cornwell, 2011) algorithm in the tclean task to account for possible deconvolution errors in wide-band imaging. In the pipeline, five rounds of phase-only self-calibration were performed before making the final image. These maps were then corrected for primary beam gain. All images used in our analysis are primary beam corrected.

The radio maps of WR 114 and WR 142 are shown in Figures 1 and 2, respectively. In these maps the positions of the WR stars and other known radio sources are marked. The details of the maps (resolution and rms) are listed in Table 2. The maps presented in the figures have been convolved with a beam size of 5.0 arcsec.

There is no radio emission detected for these stars in the observed frequency bands. Of particular mention is the field of WR 114: The Band 4 map shows faint, diffuse emission at the location of the WR star. However, this emission is most likely part of an arc-like structure associated with the supernova remnant (SNR) SNR G017.4$-$00.1 (Brogan et al., 2006; Green, 2009) in the background along the line-of-sight toward WR 114, at an estimated distance of 18.6 kpc (Pavlović et al., 2013). Other radio sources detected in both fields are marked in the respective figures. Thus, these maps give only upper limits to the flux density from the stars. From the achieved rms level of the uGMRT maps, for WR 114 we derive $3\sigma$ upper limits of 123 $\mu$Jy ($\nu L_{\nu}<4.2\times 10^{26}~{}\rm{erg\,s^{-1}}$) and 66 $\mu$Jy ($\nu L_{\nu}<3.9\times 10^{26}~{}\rm{erg\,s^{-1}}$) at 735 and 1260 MHz, respectively. For WR 142, the upper limits are 111 $\mu$Jy ($\nu L_{\nu}<2.8\times 10^{26}~{}\rm{erg\,s^{-1}}$) and 96 $\mu$Jy ($\nu L_{\nu}<4.1\times 10^{26}~{}\rm{erg\,s^{-1}}$) at 735 and 1260 MHz, respectively.

Both WR stars have been previously observed at higher frequencies. In their search for radio emission towards WR stars, Chapman et al. (1999) presented ATCA observations of WR 114. No emission was detected in none of the observed frequency bands. Both WR stars are also undetected in a VLA 6 cm observation presented by Abbott et al. (1986). In addition, Cappa et al. (2004) observed WR 114 and WR 142 at 3.6 cm as part of a VLA survey of Galactic WR stars. In their study, WR 114 is classified as a probable detection (with signal-to-noise ratio of $\approx$4), whereas WR 142 is listed as non-detection. It is to be noted that these authors advocate further confirmation of probable detections. Additionally, given the beam size ($\rm 9\,arcsec\times 6\,arcsec$) of these observations, there could be contribution from the background SNR. The 3$\sigma$ upper limits from the above studies along with the ones obtained in this work are compiled in Table 3.

We carried out a low-frequency (735 and 1260 MHz) radio continuum study of two WR stars, WR 114 and WR 142, using uGMRT observations with high angular resolution and sensitivity. At both frequencies, no radio emission is detected from either of the WR stars. Based on the non-detection, we report on 3$\sigma$ upper limits to the radio flux densities at 735 and 1260 MHz (123 and 66 $\mu$Jy for WR 114, and 111 and 96 $\mu$Jy for WR 142, respectively). These limits are discussed in the context of two scenarios: stellar wind from a single star and a CWB system.

Considering the thermal emission from the stellar wind of a single star enables us to constrain the mass-loss rate of WR 114 to $\dot{M}\,\lesssim\rm 10^{-5}\,M_{\odot}\,yr^{-1}$, which is a factor of $\sim$3 lower than the model-based value from Sander et al. (2019). In addition, our analysis does not reveal any synchrotron emission associated to a potential wind-ISM interaction. A comparison to the synchrotron-emitting WR bubble around WR 102 points to the requirement to carefully consider the properties of the direct circumstellar environment to envisage significant particle acceleration at a termination shock.

If we rather assume both WR stars to be in CWB systems, the non-detection of NT emission in the lower-frequency uGMRT maps suggests either (a) a very wide binary system (with an orbital period of several decades) not close to periastron, or (b) a close binary system with strong FFA. Alternatively, the non-detection of synchrotron emission may also indicate the lack of a companion. The latter point deserves complementary studies at other wavebands to independently constrain the multiplicity of these two WR stars.

The investigation of particle acceleration from massive stars in various configurations, either single stars or in binary/multiple systems, definitely deserves to be pushed forward with more targets investigated at various radio frequencies, at various epochs, and keeping track of all potential shock physics scenarios likely to occur in these objects. Our results show that the availability of a quite high kinetic power is not sufficient to warrant the identification of clear indicators of particle acceleration in these objects. In particular, for isolated WR stars, the efficiency of converting wind kinetic power into particle acceleration can be rather low under unfavourable conditions.

We thank the referee for valuable suggestions that helped us to improve the manuscript. This work was carried out in the framework of the PANTERA-Stars (https://www.astro.uliege.be/~debecker/pantera/) initiative. We thank the staff of the GMRT that made these observations possible. GMRT is run by the National Centre for Radio Astrophysics of the Tata Institute of Fundamental Research.

The raw data is publicly available in the GMRT webpage (https://naps.ncra.tifr.res.in/goa/data/search). The calibrated data used in this article will be shared upon reasonable request to the corresponding author.