Gaia 17bpi: An FU Ori Type Outburst
We report on the source Gaia 17bpi and identify it as a new, ongoing FU Ori type outburst, associated with a young stellar object. The optical lightcurve from Gaia exhibited a 3.5 mag rise with the source appearing to plateau in mid/late 2018. Mid-infrared observations from NEOWISE also show a 3 mag rise that occurred in two stages, with the second one coincident with the optical brightening, and the first one preceding the optical brightening by 1.5 years. We model the outburst as having started between October and December of 2014. This wavelength-dependent aspect of young star accretion-driven outbursts has never been documented before. Both the mid-infrared and the optical colors of the object become bluer as the outburst proceeds. Optical spectroscopic characteristics in the outburst phase include: a GK-type absorption spectrum, strong wind/outflow in e.g. Mgb, NaD, H, KI, OI, and CaII profiles, and detection of LiI 6707 Å. The infrared spectrum in the outburst phase is similar to that of an M-type spectrum, notably exhibiting prominent and CO (2-0) bandhead absorption in the K-band, and likely HeI wind in the Y-band. The new FU Ori source Gaia 17bpi is associated with a little-studied dark cloud in the galactic plane, located at a distance of 1.27 kpc.
Department of Astronomy, California Institute of Technology, Pasadena CA 91125 \move@AU\move@AF\@affiliationSchool of Physics, University of Exeter, Stocker Road Exeter, Devon, EX4 4QL 0000-0001-6352-5312]Sam Morrell \move@AU\move@AF\@affiliationSchool of Physics, University of Exeter, Stocker Road Exeter, Devon, EX4 4QL 0000-0002-0506-8501]Tim Naylor \move@AU\move@AF\@affiliationSchool of Physics, University of Exeter, Stocker Road Exeter, Devon, EX4 4QL 0000-0002-0631-7514]Michael A. Kuhn \move@AU\move@AF\@affiliationDepartment of Astronomy, California Institute of Technology, Pasadena CA 91125 0000-0002-0077-2305]Roc M. Cutri \move@AU\move@AF\@affiliationIPAC, California Institute of Technology, Pasadena CA 91125 0000-0001-6381-515X]Luisa M. Rebull \move@AU\move@AF\@affiliationIPAC, California Institute of Technology, Pasadena CA 91125 \move@AU\move@AF\@affiliationInstitute of Astronomy, University of Cambridge, Madingley Road, Cambridge CB3 0HA, UK \move@AU\move@AF\@affiliationCentre for Astrophysics & Planetary Science, The University of Kent, Canterbury, Kent CT2 7NH, UK \move@AU\move@AF\@affiliationJPL, 4800 Oak Grove Dr. Pasadena 91109
The accretion paradigm for young stars is reviewed by Hartmann et al. (2016). It features an early spherical-like infall from a hydrostatic core of slowly rotating molecular and dusty material, the build-up of (proto)stellar mass through a combination of the direct collapse and accretion from a disk (mediated by a fraction of the mass that is lost in accretion-driven winds/outflows), and finally, a late-accretion phase that competes with planet formation and remnant winds in completely depleting the disk. The initially high accretion rates are believed to generally decline with age. However, our physical understanding of star formation and stellar mass assembly currently relies on episodic accretion, or punctuated periods of enhanced mass accretion/outflow, in order to build up the needed stellar mass on the required time scales.
Based on a small sample of “classical” FU Ori stars (Herbig, 1977) and a few subsequently discovered FU Ori-like objects – those whose transition to the outburst state had not been observed – Hartmann & Kenyon (1996) summarized the basic scenario of episodic accretion. They envisioned that enhanced accretion would occur more frequently in more massive disks at early protostellar stages, and thus that the rate of FU Ori events would be higher for protostars, and lower for optically revealed pre-main sequence stars. Subsequent detailed modeling of instabilities arising in the inner disk (Bae et al., 2014) or the outer disk (Vorobyov & Basu, 2015) has produced quantitative predictions concerning the amplitudes, durations, and duty cycles of episodic accretion in young stellar objects.
A long-standing problem for the importance placed on the FU Ori scenario in building up stellar mass, is that FU Ori outbursts are rare. Over the past seven decades, fewer than 13 actual FU Ori outbursts have been recorded, with another 13 sources identified as FU Ori-like based on their present spectra and spectral energy distributions obtained in a hypothetical post-outburst state (Reipurth & Aspin, 2010; Connelley & Reipurth, 2018). Although the discoveries are increasing in number with time (Figure 73 in Reipurth, 2016), the rate of FU Ori outbursts remains rather poorly constrained empirically. This is an especially notable gap when compared to our state of knowledge regarding event rates for e.g. cataclysmic variable and other novae, supernovae of various classes, and even tidal disruption events.
Modern all-sky and all-hemisphere time domain surveys have the potential to better constrain the true FU Ori rate (Hillenbrand & Findeisen, 2015), and better illuminate the diversity of young star outbursts of various types, e.g. FU Ori-like vs V1647 Ori-like vs EX Lup-like (Contreras Peña et al., 2017a, b). Only when the full phase space of young stellar object variability is more completely mapped out, can we improve our understanding of how stars gain their mass.
The Gaia mission is one such photometric survey, performing repeated scans of the sky primarily directed towards establishing accurate and precise astrometry. However, changes in source brightness trigger alerts that are made publicly available via a URL interface 111Gaia Alerts; http://gsaweb.ast.cam.ac.uk/alerts. These pages are monitored by groups wishing to identify objects of interest to their science goals, and perform follow-up observations. An alert for Gaia 17bpi was issued on 2017, June 21, following a 2 mag brightness increase relative to observations from the earliest Gaia epochs. The source has continued to brighten.
While the lightcurve of Gaia 17bpi exhibits an FU Ori-like rise, several classes of large-amplitude pulsating variables (LPVs) have similarly slow rise times, as can active galactic nuclei (AGN). However, the location of Gaia 17bpi in proximity to an optically opaque cloud and region of infrared nebulosity, suggested that further investigation of the source was warranted.
In this paper we report on the environment of the new outburst source Gaia 17bpi (§2), on the characteristics of the pre-outburst object (§3), and on the substantial rise relative to previous photometry exhibited in the Gaia lightcurve as well as in NEOWISE observations, including follow-up optical photometry (§4.1-4.3). In §4.4 we present new outburst-era spectroscopy at optical and infrared wavelengths, and §5 contains a short discussion and our conclusions.
2 The Star-Forming Region Containing the New Outburst Source
Gaia 17bpi is associated with a previously faint optical (r22 mag), near-infrared, and mid-infrared point source. The position is located towards the northern end of the elongated G53.2 “infrared dark cloud”, just outside of a particularly opaque region on optical and infrared images, and adjacent to a small HII region cataloged as HRDS G053.822-00.057 (Anderson et al., 2014). The modest star-forming region is hardly studied.
The only previous relevant publication covering objects in the vicinity of Gaia 17bpi is by Kim et al. (2015). These authors illustrate the morphology of the dust (based on Bolocam) and gas (based on GRS) relative to the mid-infrared emission and absorption (see their Figure 1). Gaia 17bpi appears to be on the periphery of the cloud, beyond the CO gas contours that also delineate the optically opaque region, but between two millimeter-wavelength dust clumps. Kim et al. (2015) used Spitzer photometry to identify several hundred young stars as likely members of the star forming region which extends in a narrow filament over several degrees to the southeast of Gaia 17bpi, though not including the newly outbursting source that we report here. Kim et al. (2015) quote a kinematic distance to the cloud of 1.7 kpc.
In order to refine the distance and to assess the overall environment of the outburst source, we repeated the Kim et al. (2015) analysis by using the Spitzer/GLIMPSE (Churchwell et al., 2009), 2MASS (Cutri et al., 2003; Skrutskie et al., 2006), and Gaia DR2 (Gaia Collaboration et al., 2018) catalogs. The color criteria of Gutermuth et al. (2009) were used to identify infrared excess sources that are likely young stellar objects, then Gaia proper motions and parallaxes were used to down-select to a sample of reliable members. From infrared excess alone, there is substantial contamination from dusty giants in the background. From Gaia alone, it is difficult to identify the association due to an insufficient population of stars that would stand out in proper-motion/parallax space. As in Kim et al. (2015), Gaia 17bpi was not selected by our mid-infrared selection methods, mainly because in the available Spitzer catalog data it is detected in only the two shortest (3.6 and 4.5 m) photometry bands. The spatial distribution of our infrared excess and kinematically selected sources exhibits two clumps to the southeast of Gaia 17bpi, coincident with the morphology of the opacity in optical/infrared images. Again, Gaia 17bpi is found to be located on the periphery of the star forming region.
We assess the cloud distance based on examination of the Gaia parallax distribution of the infrared excess sources, using a weighted median approach identical to that described in Kuhn et al. (2018). Our value of pc (including both random and systematic error) is approximately 25% closer than the previous kinematic distance estimate. The total cloud mass reported by Kim et al. (2015), correcting for the distance revision, becomes a few . The cloud is comparable in its filamentary morphology and its size and mass to the Taurus dust cloud and molecular gas complex (Pineda et al., 2010). The cloud radial velocity reported by Kim et al. (2015) is km/s, which if an LSR value corresponds to .
3 The Pre-Outburst Object
3.1 SED Data Collection
The spectral energy distribution (SED) of the pre-outburst object was assembled from catalog data originating from the Gaia DR2, PanSTARRS DR1, IPHaS, 2MASS, and Spitzer/GLIMPSE-I surveys, as well as new data reduction of Spitzer and Herschel images. The results of the reductions described below are given in Table 3.1.
For the Spitzer data, the GLIMPSE catalog contains measurements only at the shortest wavelengths, 3.6 and 4.5 micron. In order to better characterize the infrared SED, we downloaded from IPAC/IRSA all Spitzer image data from IRAC (3.6, 4.5, 5.8, 8 micron) and MIPS (24, 70 micron), and all Herschel data from PACS (70, 160 microns) and SPIRE (250, 350, 500 microns). The source is apparent in all four IRAC bands, but at the longer wavelengths there is contamination from a very bright source to the northeast and only upper limits could be derived, at much higher levels than typical in less confused regions.
The position was observed at mid-infrared wavelengths as part of the GLIMPSE program (Benjamin et al. 2003; Churchwell et al. 2009) with IRAC in late 2004. The mosaics we used were assembled as part of the Spitzer Enhanced Imaging Products222SEIP; https://irsa.ipac.caltech.edu/data/SPITZER/Enhanced/SEIP/overview.html which summed up all available data from the cryogenic phase of the mission. GLIMPSE is a relatively shallow survey, but the target is clearly visible in the IRAC-1 (3.6 um) and IRAC-2 (4.5 um) bands; it is faint but present in IRAC-3 (5.8 um) and IRAC-4 (8 um). Aperture photometry was performed at the location, using an aperture of 3 native pix, with an annulus of 3-7 native pix. (A native pixel is 1.2.) Aperture corrections of 1.124, 1.127, 1.143, and 1.234 were used, as described in the IRAC Instrument Handbook333SSC; http://irsa.ipac.caltech.edu/data/SPITZER/docs/irac/iracinstrumenthandbook/. Reported errors are statistical and do not include systematic effects.
This region was also observed by MIPS at 24 microns as part of the MIPSGAL program (Carey et al., 2009) in late 2005, and mosaics re-generated as part of SEIP. However, the source was not detected. Calculating photometric limits at this location is complicated by nearby bright ISM; we performed aperture photometry at the location as if there was a source there, following suggestions from the MIPS Instrument Handbook444SSC; http://irsa.ipac.caltech.edu/data/SPITZER/docs/mips/mipsinstrumenthandbook/ that is, 5.6 aperture, annulus 5.6-10.4, and an aperture correction of 2.05.
This region was also observed by Herschel (Pilbratt et al., 2010) in late 2011 as part of the Hi-GAL program (Molinari et al., 2010) using PACS and SPIRE. The source is not visible in the Herschel High-Level Images555HHLI; http://irsa.ipac.caltech.edu/data/Herschel/HHLI/overview.html, which sum up all available data. As with MIPS-24, limits are not terribly constraining because of the bright ISM near the region. Aperture photometry obtained from the corresponding instrument handbook was performed at the location to estimate upper limits in each available band (PACS: 70 and 160 um; SPIRE: 250, 350, and 500 um).
Table 0. \Hy@raisedlink\hyper@@anchor\@currentHrefNewly Derived Infrared Photometrya for the Gaia 17bpi Progenitor
|Instrument||Wavelength (m)||Flux (Jy)||Error (Jy)|
|\tablenoteObservations were taken in 2004 as part of the Spitzer GLIMPSE and MIPSGAL programs.|
3.2 SED Analysis
The 0.5-4.5 micron photometric measurements and the longer wavelength upper limits are shown in Figure 3.2. It may be important to note that the pre-outburst photometry consists of measurements taken at different times, with the optical grizy data taken in 2012, the near-infrared JHK data taken in 1999, and the IRAC1,2,3,4 measurements taken in 2004. As these sets are non-contemporaneous, the compiled SED could be deceptive.
Photospheric models ranging in temperature from 2900-3200-3500-4000-4500 K from NextGen2 (following Hauschildt et al., 1999) result in acceptable fits to the Wien part of the SED for corresponding visual extinction values of 1-2-3-4-5 mag, respectively. Photospheres warmer than K can not be reddened enough to match the bluer wavelengths without exceeding the measurements around and beyond the SED peak. Photospheres cooler than K do not match at the bluer wavelengths, even unreddened.
Hotter temperatures minimize the near-infrared excess; however, these warmer temperatures do not match the r-band to G-band to i-band data as well as cooler temperatures do. Temperatures cooler than K imply a near-infrared excess, the amplitude of which increases as the assumed photospheric temperature decreases. The 5.8 and 8.0 micron points clearly exceed the expectations from a reddened photosphere of any temperature. Our preferred model, derived by considering a number of different normalization schemes, is the 3500 K photosphere with mag.
3.3 Other Attributes
Unfortunately we know little else about the pre-outburst object beyond its location and its colors. It was too faint for parallax or proper motion reporting in Gaia DR2, for example.
Integrating the pre-outburst photometry described above between 0.45 and 4.5 microns, and assuming the 1.27 kpc distance derived above, a source luminosity of 0.3 Lis derived for the pre-outburst object. The true value will be slightly higher once the proper reddening correction is determined. This luminosity is seemingly appropriate for a pre-main sequence star in the above temperature range. The corresponding spectral type would be mid-K to M.
In the Gaia photometry presented below, there is some indication of photometric variability in the earliest G-band measurements, taken pre-outburst. Gaia 17bpi is quoted in the Gaia DR2 catalog at mag. Given that 13 transits contribute to that measurement, the individual measurements are implied to have uncertainty 0.05 mag. However, as the uncertainty is based on the scatter between data points, any intrinsic variability will be included in the uncertainty value. Indeed most sources at this magnitude in the general area have an uncertainty of only 0.005 mag. The possible pre-outburst variability of Gaia 17bpi has an amplitude of a few tenths of a magnitude, and occurs on a timescale of a few days (Figure 4.2), consistent with “typical” T Tauri star variability (e.g. Cody & Hillenbrand, 2018; Rebull et al., 2018).
The iPHaS catalog (Barentsen et al., 2014) provides color of mag, but the very red color of mag does not make the source an obvious emitter, according to the models in Figure 2 of Barentsen et al. (2013).
The field of Gaia 17bpi was also covered as part of the UWISH2 survey (Froebrich et al., 2011). Consistent with results presented below, the pre-outburst source is detected in continuum J-band and K-band images (taken 2006-07-11 for the UKIDSS Galactic Plane Survey) but does not show an excess in an difference image ( data taken 2010-11-26). There is thus no signature of a pre-existing large-scale shocked outflow from the Gaia 17bpi progenitor, only some jets in the dark filament to the northwest of the source that are not related, and some nebulosity to the northeast that is coincident with the mid-infrared nebulosity illustrated in Figure 4.2.2. If an outflow appears over the next few years, it would be direct evidence for FU Ori outburst triggering of mass ejection.
4 THE OUTBURST OF Gaia 17bpi
At position 19:31:5.590 +18:27:52.27 (J2000.), Gaia 17bpi was flagged by an automated alerts system (Hodgkin et al., 2013, Hodgkin et al. 2019, in preparation) and announced on the Gaia Alerts public feed666http://gsaweb.ast.cam.ac.uk/alerts/alert/Gaia17bpi/ on 2017, June 23. The alert was triggered as a “delta-magnitude detection” based on the source brightening relative to earlier Gaia photometry by mag in two consecutive transits of the satellite over the source position (in March and then June, in this case). Among the stream of Gaia Alerts, this particular source was noted as being a potential young star outburst via an ongoing program to identify young stellar objects that are published on the Gaia Alerts feed. Contreras Peña et al. (2018, in preparation) present a full description of this program.
Although Gaia 17bpi was not successfully cross-matched with any specific object in the search catalogs, it was identified using what we term a “vicinity match,” that is, as being located within 2 arcmin of a confirmed or candidate young stellar object. This methodology leverages the fact that YSOs are located in spatially coherent regions of star formation, hence clustered on the sky. Further, it acknowledges that the current census information for many nearby star-forming regions remains incomplete, and uses photometric variability to flag potential YSOs that are previously uncataloged. In fact, Gaia 17bpi has five stars falling in the vicinity of the alert coordinates, given in subsection 4.1.
Table 0. \Hy@raisedlink\hyper@@anchor\@currentHrefSIMBAD-identified YSOs within 2 arcmin of Gaia 17bpi
|Name||RA (J2000)||Dec (J2000)||Separation ()|
4.2 Assembly of Lightcurve Data
Photometry from the Gaia mission was downloaded from the alerts service webpage, with the last update for our analysis occurring on 10 September, 2018. Only Gaia G-band measurements are available and error estimates are not included. We do not reproduce the numbers here, as the photometric quality should continue to improve as the mission proceeds.
The recent Gaia lightcurve was supplemented by heritage catalog data from IPHaS (Barentsen et al., 2014) and PanSTARRS (Flewelling et al., 2016), Reported r-band magnitudes were adjusted (in the case of IPHaS, first from Vega to the AB system) from the PanSTARRS AB system to Gaia’s G-band, using an empirical calibration777 We derived over the magnitude range . . Looking to other sources of recent photometry, we discovered that although there is PTF/iPTF imaging in the region spanning the time period after the PanSTARRS data point, and before/after the first Gaia detections, the source sits in a PTF chip gap. There is thus unfortunately no optical photometry recording the initial rise of Gaia 17bpi.
After the 2017 Gaia alert was issued, we monitored the object using the Liverpool Telescope. Three nights of griz photometry were obtained on 2017, June 28, October 2, and 2018, August 22. A single epoch of H-band photometry was obtained on 2017, June 28. The griz magnitudes were calibrated using Pan-STARRS AB magnitudes for a set of stars around 5 arcmin from Gaia 17bpi. The H-band data was calibrated in a similar way using 2MASS magnitudes. At the time of the first night of monitoring, the object was already 2 magnitudes brighter in both the optical and the infrared than the quiescent state as measured in earlier PanSTARRS and 2MASS data. The measurements are given in subsubsection 4.2.1.
Beginning in spring of 2018, as Gaia 17bpi began to show signs of approaching peak brightness, the Zwicky Transient Facility (ZTF) became operational. Complementing the Gaia lighturve, there are several tens of epochs of g-band and r-band photometry from ZTF taken between 2018 March through November (when the source set for the season). These data will become available with the release of the NSF/MSIP public survey part of ZTF. However, there is a single ZTF measurement from a “reference image” with and mag reported at IPAC/IRSA for a date that can not be determined at present.
Additional optical imaging was performed with the University of Kent’s Beacon Observatory (Froebrich et al., 2018) The telescope is a 17-inch Planewave Corrected Dall-Kirkham (CDK) Astrograph with a 4k 4k Peltier-cooled CCD camera and a B, V, R, I, H filter set. The final images for each night in each of the broadband filters were combined from separate integrations. Standard data reduction using bias, darks and sky-flats have been applied. BVri photometric measurements are given in subsubsection 4.2.1.
Table 0. \Hy@raisedlink\hyper@@anchor\@currentHrefOptical Photometric Follow-up During the Outburst of Gaia 17bpi
The position of Gaia 17bpi was observed by WISE (Wright et al., 2010) during its primary mission in 2010 April and October, and again twice per year between 2014 October and 2018 October during the reactivated NEOWISE survey (Mainzer et al., 2014). NEOWISE did not observe the source in 2014 April because it lies in the region of sky that was missed due to a short safe-hold during that time (Cutri et al., 2015).
Gaia 17bpi was not detected in the WISE observations, and thus it has no entry in the AllWISE Source Catalog. Brightness upper limits for the 2010 April and October observations were estimated using an 8 arcsec aperture on coadds of the 15 individual exposures covering the object in each epoch. The magnitude upper limits were computed by adding two times the flux uncertainty to the measured flux in the aperture.
The source is, however, well-detected at 3.4 and 4.6 micron in the NEOWISE observations beginning in 2015, April. The mid-infrared brightening is illustrated in Figure 4.2.2.
Because the field is complicated in the mid-infrared, due to both nebulosity and to very red nearby bright sources, the NEOWISE photometry was validated by examining individual images. Profile-fit photometric measurements from each individual exposure are available in the NEOWISE Source Database888NEOWISE; http://wise2.ipac.caltech.edu/docs/release/neowise/expsup (Cutri et al., 2015). For the NEOWISE 2015 April and later epochs, the magnitudes and uncertainties listed in Table 4.2.2 were computed from the mean and standard deviation of the mean of the Source Database profile-fit fluxes from all exposures taken during each epoch.
For the 2014, October observation set, the NEOWISE Source Database has only one reported detection with an unusually large positional offset relative to the later epochs. The source is not obvious in the individual 2014, October exposures, and we deemed the one measurement to be spurious. However the source is faintly visible at the correct position in a coadded image formed by combining the 15 individual exposures covering the position, as shown in Figure 4.2.2. Because of the complexity of the surrounding field, 3.4 and 4.6 micron photometry was performed on difference images that were constructed by subtracting coadded 2010 October images (where the source is not detected) from the coadded 2014 October images to suppress the nearby confusing objects. The flux of the source was measured in an 8 arcsec radius aperture on the difference images, which is the same sized aperture used for the “standard aperture” photometry in NEOWISE automated data processing. The aperture measurements on the difference images were put on the photometric scale of the automated profile-fit measurements from the later epochs by applying the same aperture corrections and normalization used in the automated processing (these are the w1mcor and w2mcor parameters in the NEOWISE Source Database entries). This procedure was verified by generating difference images for the later NEOWISE epochs when the source was well detected. The calibrated aperture measurements on the difference images agree with the average profile-fit measurements to within 11% in W1 and 6% in W2 in each epoch (excluding 2014 October when the source was too faint to be reliably detected in the individual exposures). The results appear in Table 4.2.2.
The assembled lightcurve includes the Spitzer/GLIMPSE-I photometry from 2004 discussed above. In addition, there is a second set of Spitzer data, taken in 2014, December, just after the first NEOWISE measurements from 2014, October. The nearby infrared dark cloud was targetted twice with IRAC 3.6 and 4.5 micron imaging, observed as part of program 10012 (PI J. Hora). Each of these observations are much deeper than GLIMPSE, with integration time 150 sec as compared to 4 sec, and the source was brighter by this time as well. Aperture photometry was performed as described earlier for the GLIMPSE re-reduction, but using aperture corrections of 1.125 and 1.120, as appropriate for post-cryogenic Spitzer data. The results appear in Table 4.2.2.
The two epochs of 2014 Spitzer data appear to have captured the outburst in progress, with the second epoch brighter than the first (by 0.4 mag at 3.6 micron and 0.5 mag at 4.5 micron), which we confirmed from visual inspection and repeated photometric measurement. The only slightly earlier 2014 NEOWISE measurement is fainter, and the first 2015 NEOWISE measurement is brighter (see Figure 4.2).
Table 0. \Hy@raisedlink\hyper@@anchor\@currentHrefNewly Measured Infrared Photometry Before and During the Outburst of Gaia 17bpi
|Telescope||MJD||3.4 m Magnitude||Error||4.6 m Magnitude||Error|
|55305.20||14.23||2 limit||12.55||2 limit|
|55487.10||13.25||2 limit||12.63||2 limit|
|57008.16||13.78 (3.6 m)||0.08||13.38 (4.5 m)||0.08|
|57014.19||13.40 (3.6 m)||0.07||12.89 (4.5 m)||0.07|
4.3 Quantitative Analysis of the Lightcurves and Color Curves
Figure 4.2 shows the assembled optical and infrared lightcurve data. Unfortunately, the beginning of the turn-up from quiescence to an early outburst stage is not well-observed in either the optical or the infrared. However, the later parts of the optical rise and the optical peak are adequately sampled, even at the low cadence of Gaia. As of 2018, September, Gaia 17bpi is 3.5 mag brighter in the optical and 3 mag brighter in the mid-infrared than it was 5-15 years ago.
The first Gaia measurements in late 2015 indicate no change in brightness from the historical optical photometry. However, about a year before, in late 2014 there is a pair of Spitzer measurements that are already 1.1 mag brighter in both I1 and I2 than the Spitzer I1 and I2 photometry from 10 years previous, in 2004. The first NEOWISE points were taken just before the two Spitzer measurements and 1.2 years earlier than the first Gaia measurement. They have large errors, and while the 3.4 micron point is not significantly brighter than the earlier Spitzer measurement, the 4.6 micron measurement could be. The outburst may have begun between October and December of 2014.
The evidence from the time series suggests that the outburst may have started in the infrared, and manifest later in the optical. As we do not sample the beginning of the burst in the optical, given the large time gap between the initial and subsequent Gaia photometry, we can not calculate the time delay, but 1.5-2 years is implied. The NEOWISE lightcurve further indicates a possible two-stage rise, in which there is perhaps 18 months of measurable mid-infrared brightening, followed by a plateau of about a year, then another at least 24 months of mid-infrared brightening corresponding to the optical rise that is well-sampled by the Gaia measurements. Figure 4.3 shows our analytic fit to the optical and mid-infrared lightcurves.
We fit the existing pre-outburst and outburst optical photometry with a sigmoid function, assuming the peak brightness occurs at the last measured epoch, which is defined as . The sigmoid is described by and our best-fit parameter values are: (meaning that the outburst starts 2=1046 days before its peak), (where the value represents a timescale for the rise, in this case a 137.4 day e-folding), and (representating the full amplitude of the rise). We note that it is not yet clear whether Gaia 17bpi has reached its peak brightness. If it continues to rise, all of the numbers quoted above would increase.
Applying the same methodology to the infrared data, we fit the first and second plateaus in the NEOWISE lightcurve separately. We find that the first rise has a 71 day e-folding time, taking 2= 593 days to rise 2.2 mag, while the second rise has a 149 day e-folding time, taking 2= 998 days to rise another 1.1 mag. The second rise in the NEOWISE data appears to correpond to the brightening in the Gaia data. Again, it is not clear whether the second infrared rise has reached its peak, so the numbers above could be lower limits.
The amplitude of the photometric brightening corresponds to a luminosity increase by a factor of 25, from the estimated 0.3 for the progenitor T Tauri star, to 7.5 in outburst.
Figure 4.2 illustrates the optical and mid-infrared color changes that have been observed. Although there is little optical color data available for Gaia 17bpi in the outburst phase (Table 4.2.1), compared to pre-outburst colors from IPHaS and PanSTARRS, the source is clearly bluer. After accounting for AB to Vega magnitude transformations, the blue-ing is about 1.2 mag in r-i color and 0.8 mag in g-r color. There should be a full g-r color time series available from ZTF when the first data release occurs. In addition, a future data release from Gaia should contain and measurements at the same epochs as the measurements.
In the infrared, the Spitzer measurements show no infrared color change between the 2004 and 2014 data. However, the first NEOWISE measurement seems to indicate a redder [3.4]-[4.6] color (a.k.a. W1-W2) compared to the historical Spitzer [3.6]-[4.5] color (a.k.a. I1-I2), by 0.45 mag. Color terms between the Spitzer and WISE filter systems can not explain the amplitude of the apparent color change; green points in Figure 4.2 show the Spitzer photometry corrected to the NEOWISE system using an empirically derived relation999 mag, with rms=0.24 mag. The NEOWISE outburst photometry is initally redder, but as the optical and second-stage-infrared burst proceeds, the mid-infrared color is observed to become bluer, by 0.4 mag. The fitted color-magnitude slope is mag with rms=0.13 mag. This slope is inconsistent with the clearing of extinction.
The evidence for significant and substantial blue-ing during the outburst, in both optical and infrared colors, points to a dramatic heating event in Gaia 17bpi.
4.4 Spectroscopic Follow-Up Data and Analysis
During the lightcurve rise, the Liverpool Telescope was used to obtain optical spectra using the low-resolution spectrograph SPRAT (R=350). The observations were carried out during the same nights as the photometric monitoring reported above. The data are displayed in Figure 4.4.
The first low resolution spectrum, obtained when the source was 2 mag brighter than quiescence, shows little in terms of emission or absorption. But as the source brightens to more than 3 mag above quiescence, distinct absorption features in NaD and appear, and strengthen over time.
The Keck/DEIMOS (Faber et al., 2003) spectrograph was then used on 2018, September 10 UT by E.N. Kirby to observe Gaia 17bpi in a 900 sec integration. The 600ZD grating provided dispersion of 0.65 Å/pix, and led to a resolution of 3.3 Å per 0.7” slit, or over 4630 to 9865 Å, with a small gap from 7211-7229 Å.
The DEIMOS spectrum (Figure 4.4) exhibits numerous strong P-Cygni-type blueshifted absorption lines, with accompanying weak redshifted emission components. These profiles are seen in NaD, H, and the CaII triplet as highlighted in Figure 4.4. Blueshifted absorption, but non-P-Cygni profiles are presented by the Mgb triplet, KI, and OI. All of the above lines indicate a strong wind with velocity several hundred km/s. The terminal velocity of the H profile, for example, is approximately km/s. LiI absorption is also present (Figure 4.4) and has some evidence for a wind component as well. The LiI equivalent width is 0.47 Å, and is notable relative to the nearby CaI 6717 Å line with =0.39 Å, though there could be minor contamination from weak FeI lines at 6703.6 and 6705.1 Å.
The H profile has =9.2 Å in blueshifted absorption with a double-trough, and then a redshifted emission component with =-1.1 Å. In H there is =5.1 Å strength in blueshifted absorption and no emission component. The CaII 8542 Å lines has =5.2 Å in its blueshifted absorption and =-1.3 Å in redshifted emission.
In addition to the lines indicating activity and youth, the optical spectrum of Gaia 17bpi also has numerous neutral species absorption lines that are typical of GK type spectra, e.g. FeI, MgI, CaI. Also like many FU Ori stars, Gaia 17bpi has a strong feature at 6497 Å that is associated with a BaII/CaI/FeI blend.
The Palomar 200” telescope and TripleSpec (Herter et al., 2008) instrument were used on 2018, September 25 UT to observe Gaia 17bpi in the YJHK bands. A total of 1 hour of integration was accumulated from 3 sets of 4 (A-B-B-A) nods. Data were reduced using a version of the XSpexTool developed by M. Cushing and W. Vacca adapted for TripleSpec at Palomar. The spectrum has resolution and is illustrated in Figure 4.4.
The classic FU Ori signatures of absorption in the 2.3 micron region and prominent absorption in the K-band and H-band are exhibited, as are absorption features due to various atomic lines. Notably HeI 10,830 Å and several HI Pa lines are apparent. Although not apparent at our resolution, these lines likely have P Cyg type wind profiles similar to those exhibited by the optical region and HI Balmer and CaII lines. We refer to the FU Ori spectral atlas of Connelley & Reipurth (2018) and specifically to their Figure 3 for a set of comparison objects.
5 Discussion and Conclusions
There is wide diversity among members of the FU Ori class regarding outburst rise times and amplitudes. Gaia 17bpi has exhibited a 3.5 mag rise in the optical over 1046 days, with a 137.4 day e-folding time, and an accompanying 3 mag brightness increase in the mid-infrared. The infrared brightening took place in two stages, the first beginning approximately 1.5 years before the optical brightening (late in 2014 and into 2015, compared to in 2017), and the second seemingly coincident with the optical brightening. Source colors show both mid-infrared and optical blue-ing during the lightcurve rise.
The unifying elements of the FU Ori class are the distinctive spectroscopic features, notably a spectral change from hotter to cooler with increasing wavelength through the optical and near-infrared, as well as the appearance of strong wind signatures that are a consequence of the onset of rapid accretion. Gaia 17bpi exhibits a GK-type absorption spectrum in the optical, and an M-type absorption spectrum in the infrared, as is typical of the FU Ori class. While the spectra are similar in the different wavelength ranges to these single-temperature spectral classes, they are not exact matches given the complexity of the “photospheres” of members of the FU Ori class. Gaia 17bpi also shows spectroscopic signatures of strong wind/outflow, and it displays the LiI 6707 Å signature of youth. These characteristics are also consistent with an FU Ori classification.
With the data in hand to date, Gaia 17bpi appears to meet the photometric and spectroscopic critera of a bona fide FU Ori source. The source thus joins V900 Mon (Thommes et al., 2011; Reipurth et al., 2012), V960 Mon (Maehara et al., 2014; Hillenbrand, 2014), [CTF93]216-2 = V2775 Ori (Caratti o Garatti et al., 2011), and HBC 722 = V2493 Cyg (Semkov et al., 2010; Miller et al., 2011) as the latest entries over the past ten years in this still-rare category. Fewer than 13 of the known 25 FU Ori objects have been observed to undergo their dramatic brightness increases, with the rest classified as such only after the fact, e.g. V582 Aur (Samus, 2009; Munari et al., 2009) was relatively recently recognized as a member of the class, in 2009, but likely outburst in the mid-1980’s (Semkov et al., 2013).
Gaia 17bpi is unique among FU Ori outbursts in having its photometric brightening detected at both optical and mid-infrared wavelengths. The burst appears to have started in the infrared, consistent with disk models that predict instabilities in the inner 0.5-1 AU of protostellar and T Tauri accretion disks as the origin of FU Ori events.
We acknowledge ESA Gaia, DPAC and the Photometric Science Alerts Team, as well as NASA’s NEOWISE team. We are extremely grateful to Evan Kirby for obtaining and reducing to 1D format the Keck/DEIMOS spectrum presented here. We are also grateful to the Palomar Observatory staff for their real-time assistance in obtaining the TripleSpec spectrum presented here. Conversations with Kevin Burdge and Sean Carey about various photometric survey data sets were beneficial in our work. The contributions of CCP and TN were funded by a Leverhulme Trust Research Project Grant, and of SM through a Science and Technology Facilities Council studentship.
Facility: Gaia, LT:IO:O, LT:IO:I, BO, Hale:DBSP, Hale:TSPEC, Keck:I:DEIMOS, IPHaS, PanSTARRS, 2MASS, Spitzer, Herschel, WISE, NEOWISE, IRSA
- Anderson et al. (2014) Anderson, L. D., Bania, T. M., Balser, D. S., et al. 2014, ApJS, 212, 1
- Bae et al. (2014) Bae, J., Hartmann, L., Zhu, Z., & Nelson, R. P. 2014, ApJ, 795, 61
- Barentsen et al. (2013) Barentsen, G., Vink, J. S., Drew, J. E., & Sale, S. E. 2013, MNRAS, 429, 1981
- Barentsen et al. (2014) Barentsen, G., Farnhill, H. J., Drew, J. E., et al. 2014, MNRAS, 444, 3230
- Caratti o Garatti et al. (2011) Caratti o Garatti, A., Garcia Lopez, R., Scholz, A., et al. 2011, A&A, 526, L1.
- Carey et al. (2009) Carey, S. J., Noriega-Crespo, A., Mizuno, D. R., et al. 2009, PASP, 121, 76
- Churchwell et al. (2009) Churchwell, E., Babler, B. L., Meade, M. R., et al. 2009, PASP, 121, 213
- Cody & Hillenbrand (2018) Cody, A. M., & Hillenbrand, L. A. 2018, AJ, 156, 71
- Connelley & Reipurth (2018) Connelley, M. S., & Reipurth, B. 2018, ApJ, 861, 145
- Contreras Peña et al. (2017a) Contreras Peña, C., Lucas, P. W., Minniti, D., et al. 2017, MNRAS, 465, 3011
- Contreras Peña et al. (2017b) Contreras Peña, C., Lucas, P. W., Kurtev, R., et al. 2017, MNRAS, 465, 3039
- Cutri et al. (2003) Cutri, R. M., Skrutskie, M. F., van Dyk, S., et al. 2003, “The IRSA 2MASS All-Sky Point Source Catalog, NASA/IPAC Infrared Science Archive.
- Cutri et al. (2015) Cutri, R. M., Mainzer, A., Conrow, T., et al. 2015, Explanatory Supplement to the NEOWISE Data Release Products.
- Faber et al. (2003) Faber, S. M., Phillips, A. C., Kibrick, R. I., et al. 2003, Proc. SPIE, 4841, 1657
- Flewelling et al. (2016) Flewelling, H. A., Magnier, E. A., Chambers, K. C., et al. 2016, arXiv:1612.05243
- Froebrich et al. (2011) Froebrich, D., Davis, C. J., Ioannidis, G., et al. 2011, MNRAS, 413, 480
- Froebrich et al. (2018) Froebrich, D., Campbell-White, J., Scholz, A., et al. 2018, MNRAS, 478, 5091
- Gaia Collaboration et al. (2018) Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al. 2018, A&A, 616, A1
- Gutermuth et al. (2009) Gutermuth, R. A., Megeath, S. T., Myers, P. C., et al. 2009, ApJS, 184, 18
- Hartmann & Kenyon (1996) Hartmann, L., & Kenyon, S. J. 1996, ARA&A, 34, 207
- Hartmann et al. (2016) Hartmann, L., Herczeg, G., & Calvet, N. 2016, ARA&A, 54, 135.
- Hauschildt et al. (1999) Hauschildt, P. H., Allard, F., Ferguson, J., Baron, E., & Alexander, D. R. 1999, ApJ, 525, 871
- Herbig (1977) Herbig, G. H. 1977, ApJ, 217, 693
- Herter et al. (2008) Herter, T. L., Henderson, C. P., Wilson, J. C., et al. 2008, Proc. SPIE, 7014, 70140X
- Hillenbrand (2014) Hillenbrand, L. 2014, ATel 6797
- Hillenbrand & Findeisen (2015) Hillenbrand, L. A., & Findeisen, K. P. 2015, ApJ, 808, 68
- Hodgkin et al. (2013) Hodgkin, S. T., Wyrzykowski, L., Blagorodnova, N., & Koposov, S. 2013, RSPTA, 371, 20120239
- Kim et al. (2015) Kim, H.-J., Koo, B.-C., & Davis, C. J. 2015, ApJ, 802, 59
- Kuhn et al. (2018) Kuhn, M. A., Hillenbrand, L. A., Sills, A., Feigelson, E. D., & Getman, K. V. 2018, arXiv:1807.02115
- Maehara et al. (2014) Maehara, H., Kojima, T., & Fujii, M. 2014, ATel, 6770
- Mainzer et al. (2014) Mainzer, A., Bauer, J., Cutri, R. M., et al. 2014, ApJ, 792, 30
- Miller et al. (2011) Miller, A. A., Hillenbrand, L. A., Covey, K. R., et al. 2011, ApJ, 730, 80
- Molinari et al. (2010) Molinari, S., Swinyard, B., Bally, J., et al. 2010, PASP, 122, 314
- Munari et al. (2009) Munari, U., Siviero, A., Ochner, P., Fiorucci, M., & Dallaporta, S. 2009, CBET 1898, 1
- Pilbratt et al. (2010) Pilbratt, G. L., Riedinger, J. R., Passvogel, T., et al. 2010, A&A, 518, L1
- Pineda et al. (2010) Pineda, J. L., Goldsmith, P. F., Chapman, N., et al. 2010, ApJ, 721, 686
- Rebull et al. (2018) Rebull, L. M., Stauffer, J. R., Cody, A. M., et al. 2018, AJ, 155, 196
- Reipurth (2016) Reipurth, B. 2016, Institute for Astronomy Special Publications, 1.
- Reipurth & Aspin (2010) Reipurth, B., & Aspin, C. 2010, in Evolution of Cosmic Objects through their Physical Activity, ed. H.A. Harutyunian, A.M. Mickaelian, and Y. Terzian. VAOA Conf. 19
- Reipurth et al. (2012) Reipurth, B., Aspin, C., & Herbig, G. H. 2012, ApJL, 748, L5
- Samus (2009) Samus, N. 2009, CBET 1896, 1
- Semkov et al. (2010) Semkov, E. H., Peneva, S. P., Munari, U., Milani, A., & Valisa, P. 2010, A&A, 523, LL3
- Semkov et al. (2013) Semkov, E. H., Peneva, S. P., Munari, U., et al. 2013, A&A, 556, A60.
- Skrutskie et al. (2006) Skrutskie, M. F., Cutri, R. M., Stiening, R., et al. 2006, AJ, 131, 1163
- Thommes et al. (2011) Thommes, J., Reipurth, B., Aspin, C., & Herbig, G. H. 2011, CBET 2795, 1
- Vorobyov & Basu (2015) Vorobyov, E. I., & Basu, S. 2015, ApJ, 805, 115
- Werner et al. (2004) Werner, M. W., Roellig, T. L., Low, F. J., et al. 2004, ApJS, 154, 1
- Wright et al. (2010) Wright, E. L., et al. 2010, AJ, 140, 1868