The TESS view of high mass X-ray binaries in the Magellanic Clouds

IRIS TASOC GSFC-ELEANOR-LITE TGLC While you are welcome to photograph this poster for your own reference, please do not share these preliminary results beyond your immediate research circle, and especially do not post online! Fig. 7. IRIS light curves for Sk -65 66. No corrections for detector variations between sectors have been applied. The TESS view of high mass X-ray binaries in the Magellanic Clouds Identifying and studying the optical counterparts of HMXBs helps us understand these potential gravitational wave progenitor systems in richer detail. The HMXBs in the Small and Large Magellanic Clouds have been extensively studied over decades, and the TESS FFIs present a prime opportunity for optical follow-up and monitoring, especially of the LMC in the southern continuous viewing zone. We present results from an ongoing study of a sample of 275 HMXBs in the Magellanic Clouds, searching for orbital periods to constrain mass loss rates, as well as possible stellar oscillations in donor stars. This work also serves as proof-of-concept for a dedicated image subtraction photometry pipeline, optimized for crowded fields and faint variable targets, making the most of the sheer breadth of data available to us in the TESS FFIs and enabling further studies of extragalactic sources. Isabel L. Colman, Ruth Angus, Michael Shara isabel.c[email protected] · ilc.fyi Work supported by NASA grant 23-ADAP23-0089 P.S. I’m faculty at NYU Liberal Studies, an interdisciplinary and undergraduate-only school, so I’m always looking for external collaborators. If you think my software or expertise could be useful for your work, please get in touch! Case study: Extreme crowding SMC X-3 aka TIC 181051846 Fig. 1. The SMC in TESS Sector 28. Fig. 4. Comparing the IRIS light curve for SMC X-3 to three other pipelines, available via MAST. See below for a discussion of contamination. While these pipelines are highly important for many other use cases, image subtraction photometry proves most suitable for a variable target in such a crowded field. For display purposes we offset the fluxes; we manually rescaled the TASOC light curve due to a difference in normalization procedures. We also removed data around a systematic at 1348 BTJD in LCs from all pipelines except GSFC-ELEANOR-LITE, which pre-removes these. Fig. 2. A 25x25 cutout around SMC X-3 in TESS Sector 1, with its position marked. SMC X-3 is a HMXB consisting of a neutron star with a 7.77s pulse period (Lazzarini+ 2019) and a Be-type donor star, orbiting at a period of 44.86d (Cowley & Schmidtke 2004). Its TESS magnitude is 14.9, and as you can see in Fig. 2 to the left, our target is surrounded by much brighter sources. We can’t yet retrieve the longer orbital period with IRIS because we still need to implement thorough stitching, but in theory longer periods can be preserved. Fig. 3. Left: A 7x7 pixel cutout around SMCX3 from Sector 1, resampled and realigned, with the target marked. Right: the weighted PRF-based mask used for photometry. Color scales are not the same. Step 1: Calculate centroids to measure overall image drift for each frame Step 2: Upsample by a factor of 10 Step 3: Realign each frame to a common center Step 4: Subtract average of all frames from each individual frame Step 5: Create weighted mask from PRF model of target Step 6: Photometry! There’s definitely a signal here, but it doesn’t look like something we expect to see from a HMXB. Could this be contamination from a nearby Cepheid? We checked with TESS_Localize (Higgins & Bell 2018) but the results are inconclusive. We also compare our results to single-pixel and 3x3-pixel simple aperture photometry to check signal dilution, and we animate the difference images visualize the strongest sources of variable signal. Combining these checks with the way aperture masks are constructed in the IRIS pipeline, we can say this signal is accurate to within one pixel. But in this crowded region of the SMC, there are typically 7–12 Gaia sources within each TESS pixel. The case of SMC X-3 illustrates both the opportunities and challenges presented by this project. Image subtraction photometry is capable of probing even the densest regions of the sky, and can produce results at faint magnitudes provided the variable signal is strong enough: most of the Gaia sources in this particular pixel are between 17 and 20 G magnitude. Future work will produce a catalog of light curves for variable sources in the Magellanic Clouds, and cases like this provide vital training data for formulating contamination fractions and performing accurate classification. th th Case study: Stellar oscillations SK -65 66 Fig. 6. The LMC in TESS Sector 2. Sk -65 66 is a HMXB (Haberl+ 1995) with its donor star confirmed to be a B0 supergiant (Negueruela & Coe 2002). There are no measured orbital periods or NS pulsations for this system. There is also no TIC entry, but nevertheless we have a well-characterized optical counterpart, with a Gaia magnitude of 13.0. Unlike SMC X-3, Sk -65 66 is in a less-dense region of sky. The LMC in general also has a lot more TESS coverage than the SMC, owing to its proximity to the southern CVZ. So far we’ve analyzed Sk -65 66 in the first year of the TESS mission, where it was covered by Sectors 1, 2, and 4–12. We excluded sectors 4 and 11 from the present analysis due to data peculiarities that will be handled in future versions of the code. Bibliography Aerts et al., K2 photometry and HERMES spectroscopy of the blue supergiant ρ Leo: rotational wind modulation and low-frequency waves, MNRAS, 2018 Colman et al., The Kepler IRIS Catalog: Image Subtraction Light Curves for 9150 Stars in and around the Open Clusters NGC 6791 and NGC 6819, ApJS, 2022 Cowley & Schmidtke, The Orbital Period of SMC X-3 from Optical Photometry, AJ, 2004 Green et al.,15 000 ellipsoidal binary candidates in TESS: Orbital periods, binary fraction, and tertiary companions, MNRAS, 2023 Haberl et al., RX J0532.5-6551: a new high mass X-ray binary in the LMC, A&A, 1995 Higgins & Bell, Localizing Sources of Variability in Crowded TESS Photometry, AJ, 2018 Koumiotis et al., Variability of Galactic blue supergiants observed with TESS, A&A, 2025 Lamb et al., The Runaways and Isolated O-Type Star Spectroscopic Survey of the SMC (RIOTS4), ApJ, 2016 Lazzarini et al., Neutron Stars and Black Holes in the Small Magellanic Cloud: The SMC NuSTAR Legacy Survey, ApJ, 2019 Lucke et al., Discovery of X-ray pulsations in SMC X-1, ApJ, 1976 Ma et al., Variability of Blue Supergiants in the LMC with TESS, ApJ, 2024 Negueruela & Coe, The population of massive X-ray binaries. I. The Large Magellanic Cloud, A&A, 2002 Spejcher et al., An Investigation into the Variability of Luminous Blue Variable Stars with TESS, AJ, 2025 van Paradijs & Kuiper, Optical photometry of massive X-ray binaries: SMC X-1/Sk 160, A&A, 1984 Other TESS FFI pipelines: GSFC-ELEANOR-LITE: Powell et al., The NASA GSFC TESS Full Frame Image Light Curve Data Set, RNAAS, 2022 TASOC: Handberg et al., TESS Data for Asteroseismology: Photometry, AJ, 2021; Lund et al., TESS Data for Asteroseismology: Light-curve Systematics Correction, ApJS, 2021 TGLC: Han & Brandt, TESS-Gaia Light Curve: A PSF-based TESS FFI Light-curve Product, AJ, 2023 QLP: Huang et al., Photometry of 10 Million Stars from the First Two Years of TESS Full Frame Images: Part I, RNAAS, 2022 Python packages used for this work: AstroPy, Astroquery, Lightkurve, Matplotlib, NumPy, Pandas, Reproject, ScikitImage, SciPy, TESS_PRF We detect a peak at 3.427 μHz (0.297 cd ) and rule out contamination from a nearby bright M giant with known long period variability. The observed variability is comparable to light curves for blue supergiants found in K2 (Aerts+ 2018) and TESS (particularly the LMC targets covered by Ma+ 2024, and see also Spejcher+ 2025, Koumiotis+ 2025). Further study and modeling will be needed to gain a thorough understanding of this system. In the meantime, -1 Sk -65 66 serves as an example of the exciting new discoveries that have already resulted from studying HMXBs in the Magellanic Clouds with IRIS photometry. The IRIS pipeline IRIS stands for “increased resolution image subtraction,” a variation on classical image subtraction photometry that was designed for use on the Kepler superstamp images of open clusters NGC 6791 and 6819 (Colman+ 2022). Since the publication of the original IRIS pipeline, we have made significant improvements in the course of adapting the code to be used on TESS data. The principle behind image subtraction (also often called difference imaging) is that subtracting an average image before performing photometry will treat background flux and all non-variable sources as noise. The resulting images will only contain variable flux. But to accurately generate and subtract an average image, we must first realign all images in the time series to remove the effects of detector drift. We use the centroid measurements of bright sources in the target’s immediate area to guide realignment, and the increased resolution allows for finer interpolation when realigning each frame. One exciting feature of IRIS is that our implementation of motion correction removes the scattered light signals in TESS data. The IRIS light curves that you see on this poster have been minimally processed: for display purposes, long-term trends (which may yet be significant!) have been removed by a simple linear fit. Otherwise, this is the raw photometric data. No fancy correction necessary! An eventual goal of this work is to use IRIS to produce light curves for 100,000+ sources in the LMC and 10,000+ in the SMC. Background & Motivation High mass X-ray binaries (HMXBs) are characterised by their hard X-ray luminosity and extreme variability. The accretor is a compact object, typically a neutron star, though some systems have black hole accretors and a handful are thought to contain white dwarfs. The donor star is a Bor O-type star, frequently a Be star. Some systems, such as SMC X-1 (below) and Sk -65 66 (right), contain supergiant donors. Mass transfer in HMXBs results in orbital decay over long timescales, necessitating long term monitoring of these systems. The mass transfer process is crucial to understanding the evolution of high-mass binary systems in general, which are a key pathway to producing binary compact objects, gravitational wave (GW) event progenitor systems. Ultimately, studying HMXBs will improve rate estimates of compact object mergers leading to GW events. Thanks to a long lineage of thorough observations in both the X-ray and optical domains, the population of HMXBs in the Magellanic Clouds has been wellstudied over decades, and many of these targets have known orbital ephemera and observations of intrinsic stellar variability. The short cadence and broad coverage of TESS FFI observations presents an opportunity for both follow-up and new discoveries. The Magellanic HMXBs also provide a small and wellconstrained sample of targets for a pilot study that will help us calibrate a photometric pipeline for specific use in extremely crowded regions of the sky, opening up a host of possibilities in extragalactic astronomy with TESS. Case study: Orbital periods SMC X-1 aka TIC 426012850 SMC X-1 is one of the most famous HMXBs. It consists of a B0 supergiant (Lamb+ 2016) and a neutron star with a pulse period of 0.72s (Lucke+ 1976). In both Xray and optical light curves, it has been observed to show variation corresponding to the orbital period of the system, well-constrained to 3.89d (e.g. van Paradijs & Kuiper 1984). The X-ray variability shows eclipses. The optical light curve looks like a typical ellipsoidal binary, but in this case the variability is due to X-ray heating that causes tidal-like distortion in the supergiant. This is a key phenomenon through which we expect to identify orbital periods for HMXBs in this study. There are high-quality light curves for SMC X-1 available from the QLP, TASOC, TGLC, and GSFC-ELEANOR-LITE pipelines, and the 3.89d period is also retrieved by the Green+ 2023 survey of ellipsoidal variation in TESS. So this is not a new result, but it is a nice sanity check! Fig. 5. IRIS light curves for SMC X-1 in Sectors 27 and 28.