ARCSAT ID NUMBER: AS07

DESCRIPTIVE TITLE: Confirming Galactic, M31, and M33 Candidate LBVs Through Photometric Variability

PI: Guy Stringfellow

OBSERVER(S): Guy Stringfellow plus 5 CU undergraduate students

UNCERTIFIED/UNTRAINED OBSERVERS: 
Guy Stringfellow and the 5 CU undergraduate students
Natalie Bremer   CU APS undergraduate
Sara Grandone  CU APS undergraduate
Kat Shanahan   CU APS undergraduate
Jacob Wilson   CU APS undergraduate
Jesse Witbrod   CU APS undergraduate


COLLABORATORS:
 
CONTACT INFORMATION: (PI/OBSERVER email, phone)
Guy Stringfellow    Guy.Stringfellow at colorado.edu     cell 303-506-3160

TIME REQUESTED: 
Run 1:   June 23-29 (alternate June 30 - July 6)   [on-site?]
Run 2:   August 25-31  (alternate August 18-24)
Run 3:   September 22-28 (alternate Sept 29-Oct 5)

INSTRUMENT:  SurveyCam 

FILTERS: 
MSSSO VRI, WC H-alpha, Sloan gri

COMMENTS: 
We may all come down to APO to conduct one of the observing runs on-site.
To be arranged if APO space and student constraints allow.

BRIEF SCIENCE JUSTIFICATION:  

Luminous Blue Variable stars (LBVs)  represent an extremely rare class of stars that define an  important and poorly understood phase in the evolution of the most massive and most luminous stars. The LBV phase lasts less than 10,000 years, and the evolutionary progression involving the LBV phase and the stars final subsequent outcome are not clear. For example, are all (or any) LBVs derived from a post red supergiant phase? Does the LBV phase invariably precede and lead to Wolf-Rayet stars?  Evidence now suggests that the LBV phase can be the final endpoint for some massive stars, leading directly to the explosion of Type II supernovae: SN 2011ht (Fraser et al. 2013, ApJ, 779, 8) and SN 2009ip (Pastorello et al. 2013, ApJ, 767, 1; Mauerhan et al. 2013, MNRAS, 430, 1801). But there is yet another fundamental problem associated with the post-main-sequence evolution of massive stars. It has become apparent that main-sequence mass loss rates for the most massive stars have previously been overestimated, with rates revised downwards by factors of 5 or more. With these revised rates, massive stars can not lose enough of their mass through winds during the main-sequence and early post-main-sequence evolution to reconcile observations. This indicates that much of the stellar mass needs to be lost in some other way. This places more importance on the major eruptive mass-loss events defining the LBV phase. The LBV phase may now be the critical phase where most of the star's mass-loss must now occur prior to stars becoming WR stars or exploding as supernovae. 

Despite the ever increasing importance of the critical LBV phase, there has been an extreme paucity of confirmed LBVs known in the Galaxy. Prior to 2010 there were only about a dozen confirmed Galactic LBV stars known (Clark et al. 2005, AA, 435, 239; Vink 2012, ASSL, 384, 221). Amongst these are the prototypes P Cygni and eta Carinae. Such paucity has previously been used to argue against the LBV phase as being a statistically important evolutionary phase in the lives of massive stars. The discovery of massive nebular shells in the Spitzer and WISE archival surveys are believed to have been produced by massive evolved stars,. The progenitor stars of these shells have eluded  detection despite dedicated surveys and techniques designed to discover them in the optical and near-IR. We have successively identified dozens of new massive evolved stars that spectroscopically resemble LBVs and related WN transitional stars (Stringfellow et al. 2012, IAU Symp. 282, p.267; Stringfellow et al. 2013, ASP Vol. 465, 514; Gvaramadze et al. 2012, MNRAS, 421, 3325; Burgemeister et al. 2013, MNRAS, 429, 3305). These stars are hiding within their highly extincted local environments, with their associated nebulae acting as signposts as to their location. The deep optical imaging we are conducting of these stars are revealing optical detections (e.g., Stringfellow et al. 2013), and the spectral and photometric monitoring for several are confirming them indeed to be bona fide LBVs. 

Meanwhile, others have been conducting searches for elusive extragalactic LBVs in local group galaxies by performing deep optical imaging surveys with followup spectroscopy of promising massive star candidates (e.g., Massey et al. 2007, AJ, 134, 2474). Only a few confirmed extragalactic LBVs (typically the brightest ones or those that have displayed dramatic variability, signaling interest) were previously known: for example, 4 LBVs were known to reside in M31, 5 in M33, and 3 in M101. Recent optical spectroscopy efforts have revealed an additional 20 new candidate LBVs in M31, and 35 in M33 (Massey et al. 2007). These new candidate LBVs typically range from 20 < V < 16. To confirm these new extragalactic candidate LBVs as bona fide LBVs, they must be shown to undergo either spectroscopic or photometric variability. As spectra requires hours of integration time with large aperture telescopes, and even then can render low S/N spectra, the best approach to confirming these as LBVs is to demonstrating photometric variability through monitoring. LBV variability has been demonstrated to occur on time scales ranging from months to years, which is tractable with a monitoring program with cadence times of order a month. The proposed monitoring will likely yield other transient events, including novae, which will also be studied. It should also be noted that we do not understand how the known, confirmed LBVs in fact behave; just how variable are they? Var C in M33 was recently noticed to be undergoing an enhanced (eruptive) phase (AstroTel 5403), while Romano's star in M33 is currently in a deep minimum state (AstroTel 5846).  We have been monitoring the known Galactic LBVs already, but the proposed program will extend these efforts to the nearby known extragalactic population as well. This will lead to a better overall understanding of the behavior of LBVs, and how similar or not the Galactic and extragalactic populations are regarding the
frequency and amplitude of their respective variability (or quiescence!).

Our Galactic targets lie between RA=17-19 hours scattered around and in the Galactic plane (Dec=+/-24 deg). Thus, they are first half of the night targets during the May-Oct ARCSAT scheduling period.  M31 has RA=00h44m and Dec=+41deg, M33 has RA=01h33m and Dec=+30deg, which are both up during the 2nd half of the night during the early scheduling period. Both the Galactic plane targets are setting in September when M31 and M33 are up essentially the entire night. We request three one-week observing gray-dark runs scheduled one each in June, August, and Sept-Oct, with each separated by roughly 1 month. Imaging conducted on the WIYN 0.9m at KPNO shows that under average sky conditions a 200s exposure in V yields photometry at 19 mag better than 10%. To mitigate cosmic rays, NEOs, and potential variable detector pixels/noise artifacts, 3 exposures are done in series in each filter that also increases the overall S/N of each field. Compensating for the smaller aperture of the ARCSAT 0.5m, we estimate that 3x300s exposures in the broad band filters should suffice. This will be tested and an observing protocol established during the first observing run. As the expected (long-term) variability expected for the LBVs is of order 1 mag, a precision of 0.1-0.2 mag for the faintest targets (or during the faintest minima) suffices to demonstrate significant variability. We also wish to test out deep H-alpha imaging. Due to the narrow band-width these exposures will need to be longer - typically 10-20 minutes each to be sky limited. Of the several 2x2 H-alpha filters that could possibly be used, the the WC H-alpha filter with a 50A width is likely the best choice. The CU H-alpha filter is 3"x3" and apparently has AR coating degradation, so this may not be usable with ARCSAT. Though the extragalactic candidate LBVs are spread over the extended galaxies in M31 and M33, planning indicates we only need about 8 tiles, with each tile representing the ARCSAT 0.9m field-of-view. M33 will require more tiles than M31. Photometry will be conducted in 2 filters (MSSSO V and either R or I) plus some additional H-alpha imaging as time allows. It would be nice to also have the Sloan gri filters in to explore their use and efficiency. The reason for using MSSSO VRI filters are for comparison with previous photometry; in particular with Massey et al. 2007. 

I have a team of 5 undergraduate students working with me this summer, who will participate fully in all aspects of this program. As half have been working with me over the last year, they have already been trained in CCD imaging reduction, and are presently learning photometry reduction skills. Several observed last year in our WIYN 0.9m program at KPNO, and on the CU campus 0.5m SBO telescope, so are familiar with observing techniques and protocol. We may come down to APO to conduct one (June?) observing run on site, if space at APO allows and summer class schedules for a few students permits.