Investigations into Infrared-Selected Variable Stars

Abstract

The diversity seen in the infrared variable sky is vast, and understudied when compared to variability in the optical regime. Utilising near-infrared (NIR) colours and amplitudes can be a powerful tool in the classification of the various members of this population, made easier in combination with spectroscopy. Understanding the demographics of NIR variables is key when considering pre-main sequence stellar evolution, because high-amplitude variability in YSOs can trace eruptive behaviour, which in turn is crucial in understanding the ’protostellar luminosity spread problem’. Eruptive YSOs were traditionally thought to encompass two categories, FU Orionis (FUors) and EX Lupi (EXors), the former having higher amplitude bursts on the order of centuries, and the latter with smaller, possibly repeating events on time-frames between months and years. More recent work has established a much broader zoo of observable eruptive behaviours, which warrant further investigation. For example eruptive events lead to rapid increases in the temperature of the disk environment and thus drastically alter the chemical conditions within the disk, which would then impact planet formation (Cieza et al., 2016, see). Using the widely spaced epochs of the UKIDSS Galactic Plane Survey (UGPS), in concert with the 2 Micron All-Sky Survey (2MASS), stellar variability can be detected on a variety of timescales. Probing the long term behaviour of these variables can be accomplished with the ≈20 epochs (covering 12 years at present) of MIR photometry from the WISE and NEOWISE missions, and thus can be used to identify candidate eruptive variable YSOs. In so doing both the rates of eruptive behaviour in class I and class II YSOs (younger and older systems respectively), and the overall demographics of variability in these systems can be inferred. We used a combination of K-band spectroscopy and long-duration mid-infrared light curves to determine the classification of 29 ’likely’ YSOs (and 1 candidate AGB star). These were selected on the basis of confirmed large amplitude K-band variability in addition to a projected membership of a local star forming region. We find 2 stars display features consistent with a FUor classification, 1 to have an EXor like spectrum, likely associated with a long duration outburst, and 6 other young stars with emission spectra and large outbursts (and thus could be EXors in quiescence). In addition to these candidate eruptive variables (EVs), we also find 4 ’dippers’ with a range of variability timescales, from <1 year to >10 years. Finally we note that 4 stars are now of uncertain classification, after their spectra and near-infrared colours showed similarity to D-type symbiotic stars (albeit only two have clear 13CO absorption lines, which are associated with this class of evolved star). Following this we construct an expectation value for the prevalence of eruptive variables within a sample of YSOs, given previously observed high amplitude variability, finding a value of 14+7−2%. This compares to the ≈10% found for YSOs more generally (from a sample of known Spitzer colour selected YSOs - see below), and thus demonstrates that known variable YSOs may be more likely to be eruptive. The value of eruptive candidates then rises if we include 6 further EXor candidates, which lack recent eruptions. We then examine a less-biased sample, again using both UGPS photometry and NEOWISE time-domain light-curves to identify further EV candidates. This sample was constructed from the sample of Cygnus-X YSOs from Kryukova et al. (2014), and includes all stars in the the above catalogue that have a single UGPS detection and a NEOWISE or unWISE counterpart. The resulting analysis provided an expectation value for EVs of 10+2−1%, in line with previous predictions. Two unexpected results from this work were the lack of detection of the long duration ’dippers’ (as mentioned earlier), and the discovery of 5 YSOs with high amplitude, long duration bursts (on the order of a decade). Investigating some of the unclassified (or mis-classified) members of the Lucas et al. (2017) sample, led to the discovery of a new short-period intermediate polar (IP), of novel type. We utilised a thorough multi-wavelength observing strategy covering x-ray imaging, optical spectroscopy, and multi-colour optical photometric monitoring. This combined with existing archive data from IGAPS, PANNSTARRS and Gaia, led us to conclude that we had discovered a new IP with a NIR-bright SED, likely a result of a very complete disk, as can be interpreted from the H I line ratios. Finally, three stars with NIR/WISE colours that were neither entirely YSO, nor AGB-like were analysed with a range of archival and VVV/X data, to place an estimate on their true classification. Three independent distance methods were trialled on one of these stars (GPSV3 from Contreras Pe˜na et al. (2014a)), which allowed us to make the conclusion that the star is likely some form of O-rich AGB star in the galactic halo. The three distance measures trialled were as follows: Mira period-luminosity relation derived fluxes used in an empirical AGB flux-distance relation, NIR-fitted proper motion based kinematic distances, and spectroscopically confirmed radial velocity distances. Applying the same Ishihara et al. (2011a) empirical measurements to the other two stars (GPSV15 & GPSV34), gives a similar classification, which given the similarity of their light curves could be expected.