H{\alpha} imaging for BeXRBs in the Small Magellanic Cloud

TL;DR

This study used H-alpha imaging to identify and analyze Be/X-ray binaries in the Small Magellanic Cloud, providing insights into their population and formation rate in a low-metallicity environment.

Contribution

It presents a large-scale H-alpha survey of the SMC to identify Be/X-ray binary counterparts and quantify their fraction among OBe stars, offering new data on their formation rate.

Findings

Identified ~10,000 H-alpha emission sources in the SMC.

Found the OBe/OB star fraction to be 13%.

Estimated BeXRBs/OBe ratio to be 0.002-0.025.

Abstract

The Small Magellanic Cloud (SMC) hosts a large number of high-mass X-ray binaries, and in particular of Be/X-ray Binaries (BeXRBs; neutron stars orbiting OBe-type stars), offering a unique laboratory to address the effect of metalicity. One key property of their optical companion is H{\alpha} in emission, which makes them bright sources when observed through a narrow-band H{\alpha} filter. We performed a survey of the SMC Bar and Wing regions using wide-field cameras (WFI@MPG/ESO and MOSAIC@CTIO/Blanco) in order to identify the counterparts of the sources detected in our XMM-Newton survey of the same area. We obtained broad-band R and narrow-band H{\alpha} photometry, and identified ~10000 H{\alpha} emission sources down to a sensitivity limit of 18.7 mag (equivalent to ~B8 type Main Sequence stars). We find the fraction of OBe/OB stars to be 13% down to this limit, and by investigating this fraction as a function of the brightness of the stars we deduce that H{\alpha} excess peaks at the O9-B2 spectral range. Using the most up-to-date numbers of SMC BeXRBs we find their fraction over their parent population to be ~0.002-0.025 BeXRBs/OBe, a direct measurement of their formation rate.

Full text

††volume: xxx††journal: Title of your IAU Symposium††editors: A.C. Editor, B.D. Editor & C.E. Editor, eds.

1 Introduction

The Small Magellanic Cloud (SMC) has been a major target for X-ray surveys due to our ability to detect sources down to non-outbursting X-ray luminosities $(L_{X}\sim 10^{33}\,\textrm{erg s}^{-1})$ and its impressive larger number of High-Mass X-ray Binaries (HMXBs; [Haberl & Sturm 2016]). However, the X-ray properties alone cannot fully characterize the nature of each source. HMXBs consist of an early-type (OB) massive star and a compact object (neutron star or black hole), which accretes matter from the massive star either through strong stellar winds and/or Roche-lobe overflow in supergiant systems or through an equatorial decretion disk in, non-supergiant, OBe stars (Be/X-ray Binaries; BeXRBs). The compact object dominates the X-ray spectrum while the companion dominates the optical spectrum. Thus, to understand the nature of BeXRBs we need to study their optical counterparts, which should be consistent with OBe stars. These are massive stars that show Balmer lines in emission, of which H$\alpha$ is typically the most prominent. Although the SMC is close enough to resolve its stellar population, we still lack the identification of the optical counterparts or their optical spectral classification for a large fraction ($\sim 40\%$ of the candidate HMXBs) of the most recent census (121 candidates in total; [Haberl & Sturm 2016]). To address this issue we take advantage of the fact that OBe stars display H$\alpha$ in emission, making them easily discernible from other stars in H$\alpha$ narrow-band images, and we performed a wide H$\alpha$ imaging survey of the SMC to reveal prime candidates for BeXRB optical counterparts.

2 Observations and Data Reduction

We used the Wide Field Imager (WFI@MPG/ESO 2.2m, La Silla, on 16/17 November, 2011) and the MOSAIC camera (@CTIO/Blanco 4m, Cerro Tololo, on 15/16 December, 2011) to observe 6 and 7 fields in the SMC, respectively. Given their large field-of-views ($\sim 33^{\prime}\times 33^{\prime}$) we covered almost the whole galaxy. Each field was observed in the $R$ broad-band (the continuum) and H$\alpha$ narrow-band filters. A dithering approach was needed to cover camera chip gaps, and the exposure time was selected to achieve a similar depth ($R\sim 23$ mag) in both campaigns to allow for coverage of late B-type stars at the distance of the SMC. Additionally, a set of spectrophotometric standards was observed to flux calibrate the results. Theli111 http://www.astro.uni-bonn.de/~theli/ was used to reduce and produce the final mosaics from the WFI data. For the MOSAIC data we retrieved the reduced data products from the NOAO online pipeline222http://portal-nvo.noao.edu/search/query and then combined them using Iraf’s mscred package. We finally re-sampled (with Swarp333http://www.astromatic.net/software/swarp) the mosaic images for each field, using a common center and frame size (for details see [Maravelias 2014]).

Because of the high source density we performed PSF photometry with Iraf’s daophot by properly selecting its parameters for each field. However, we ran the source detection on the broad-band $R$ image only, as the same process in H$\alpha$ would result to many spurious sources due to the Hii regions of the SMC. This ($R$-selected) source list was used to perform photometry on H$\alpha$. We first screened the (flux-calibrated) daophot results to select stellar sources, according to their $\chi^{2}$ ($\sim 1$) and sharpness ($|\textrm{sharp}|<0.5$) values. Since we were interested in OB stars we kept sources brighter than $R=18.7$ mag, which corresponds to B8 spectral-type stars at the distance of the SMC. We cross-correlated the two filters ($R$ and H$\alpha$) to identify the common sources and then with the MCPS catalog ([Zarithsly et al. 2002]) to obtain their $V,B$ photometry. Using the locus of OB stars ([Antoniou et al. 2009]) we selected the best OB candidate sources, for which we calculated their (H$\alpha-R$) index, its error, and SNR (following [Zhao et al. 2005]).

Since the $R$ filter includes the H$\alpha$ region the corresponding baseline for stars without any H$\alpha$ excess would be equal to (H$\alpha-R$)=0 mag. However, due to the differences between the two filters and the range of spectral types considered, this is not 0 (see [Maravelias 2014]). To overcome this we define the baseline (H$\alpha-R$) value for non-H$\alpha$ excess stars individually for each field based on the mode ($\langle\textrm{H}\alpha-R\rangle$) and standard deviation ($\sigma$) of the (H$\alpha-R$) distribution of all OB stars in each field. We consider as best H$\alpha$ emitting candidates the sources with: $(\textrm{H}\alpha-R)<\langle\textrm{H}\alpha-R\rangle-5\times\sigma$, and SNR$>5$.

3 Results and Discussion

Our survey reveals 9808 H$\alpha$ emitting sources in the SMC. This is 2 to 4 times more sources from other previous surveys (1844 sources; [Meyssonnier & Azzopardi 1993]), mostly due to our deeper coverage down to $V$=18.5 mag instead of $R\sim$16.5 mag.

From our analysis we know the number of OB stars and the corresponding number of emission-line stars (i.e. OBe). This allows us to derive the OBe/OB fraction for each field. We find an average value of $\sim 13.3\%$ across the SMC, consistent with previous studies (e.g. $\sim 5-11\%$ from [Iqbal & Keller 2013]). This fraction is only a lower limit of the actual population of the OBe stars since their activity is a transient phenomenon and only a fraction of them is active in a certain epoch. Furthermore, if we examine the relation of this fraction with $R$ magnitude ($1\,\textrm{mag}\sim 3$ spectral sub-types at the distance of the SMC), we notice a peak at $\sim$15 mag (corresponding to O9-B2) and a fast drop with magnitude (equal to later spectral types). This trend is consistent with previous observations, but we extend it to later B-type stars (from [Martayan et al. 2010]: peak at B2 but limited to $\sim$B3). Moreover, it confirms theoretical models that predict a peak of that ratio at B3, as a result of the critical rotational velocity ([Maeder & Meynet 2000]).

Given the numbers of OBe stars and BeXRBs we derive the BeXRBs/OBe fraction in the range $\sim 0.002-0.025$, which provides us with the formation efficiency of these systems with respect to their parent population. This is a direct measurement of their formation rate, which can place constraints on stellar population synthesis models (e.g. [Belczynski et al. 2008]). Currently, we are working on the cross-correlation of this catalog with the most recent list of candidate BeXRBs in the SMC ([Haberl & Sturm 2016]) in order to identify more optical counterparts.

Acknowledgements: GA ČR (14-21373S); RVO:67985815; NASA Grant NNX10AH47G; The State Scholarships Foundation of Greece (IKY); IAU travel grant.