Outbursting Young Stellar Object PGIR 20dci in the Perseus Arm

Episodic accretion in young stellar objects is an important aspect of mass accumulation in stars. It also plays a determining role in the evolution of circumstellar disks and likely influences the process of planet formation, as well as planetary survival. Armitage (2019) provides a review of disk physics in young stars, including nonsteady accretion and disk instabilities.

In "classical" T Tauri–type young stars, the accretion process is controlled mainly by physics in the inner disk and magnetospheric region that connects the disk to the star. Observed time variability in broadband photometry and in emission-line profiles of these sources is dominated by stochastic behavior, with timescales from hours to days (e.g., Costigan et al. 2014; Findeisen et al. 2015; Cody & Hillenbrand 2018). The empirical zoo of accretion-related photometric behavior has been measured in unprecedented detail for young stars by space missions such as CoRoT, Spitzer, Kepler, and now TESS, all of which sample timescales from minutes to a few months. Ground-based surveys such as the All-Sky Automated Survey for SuperNovae (ASAS-SN; Shappee et al. 2014), Palomar Transient Factory (PTF; Law et al. 2009), Zwicky Transient Facility (ZTF; Bellm et al. 2019; Graham et al. 2019), and Asteroid Terrestrial-impact Last Alert System (ATLAS; Tonry et al. 2018) in the optical and VVV (Minniti et al. 2010) in the infrared, by contrast, typically have lower precision and cadence (1 day to a few days) but the advantage of long-duration monitoring. Gaia and WISE+NEOWISE also provide precise space-based photometric data, but at much lower cadence (months), with total duration now exceeding a decade. Many of these surveys have all-hemisphere or even all-sky coverage, enabling rare event detection—including episodic accretion outbursts in very young stars.

Accretion-driven outbursts associated with young stellar objects have been identified at an increasing rate over the past decade. The discoveries have come from both the ground-based and the spaced-based platforms. Infrared surveys have the advantage of being able to penetrate the high levels of extinction toward the molecular clouds where young stars are born, and to some extent the additional circumstellar extinction of the youngest self-embedded sources.

Although young star outbursts appear to come in a continuum of amplitudes and rise times, for historical reasons (Herbig 1977), the labels EX Lup–type bursts (also known as EXors) and FU Ori–type outbursts (also known as FUors) are often applied. The former are more frequent, characterized by brightness increases of a few magnitudes in amplitude, lasting a few months to 1–2 yr, and have emission-line spectra. The latter are more infrequent and can be up to 5–6 mag in amplitude, lasting decades to centuries, and with absorption-dominated spectra except for P Cygni structure in some wind lines. An increasing number of objects are unable to be cast neatly into either of these categories, however. Our knowledge of the spectrum of amplitudes, durations, and duty cycles for young star outbursts is still in its infancy.

The large-scale FU Ori outburst events are intrinsically rare. Furthermore, those few that have been captured in progress have not been particularly well characterized. ⁸ The most recent FU Ori outburst discoveries are improving the situation, e.g., Gaia 17bpi (Hillenbrand et al. 2018), Gaia 18dvy (Szegedi-Elek et al. 2020), and NWISE-F J213723.5+665145 (Connelley & Reipurth 2020; Stecklum 2020). The total census of FU Ori–type stars numbers <30 objects at present, with less than half of these having had their outburst actually observed—rather than inferred from imaging at well-separated pre-outburst and post-outburst epochs.

We report here on the discovery of substantial photometric brightening of an object designated as PGIR 20dci. This source is located at R.A. = 00:52:20.21, decl. = +56:34:03.9, and we associate it with the NGC 281-W star-forming region. The object appears to be a bona fide FU Ori star that has just reached a plateau phase in its brightening.

NGC 281, also known as Sh 2-184, is an H ii region located in the Perseus spiral arm. While the traditional distance in the literature on the stellar population is ∼2.2 kpc, astrometry from Gaia DR2 (Gaia Collaboration et al. 2018) indicates a somewhat farther 2.8 kpc. This is also the distance derived by Sato et al. (2008) from parallax studies at radio wavelengths of a maser population associated with NGC 281-W (where PGIR 20dci is located). These results place the region on the far side of the Perseus arm, and about 300 pc out of the Galactic plane.

IC 1590 (<2 Myr) is the main stellar cluster within the nebular region, and contains the bright ionizing sources collectively designated as HD 5005. This source is now a spatially resolved multiple-star system, consisting of components with spectral types O6.5, O8, and O9 (Guetter & Turner 1997). The cluster thus draws analogies in the literature as a Trapezium-type system. The stellar population of the IC 1590 cluster was investigated by Guetter & Turner (1997) and Sharma et al. (2012), the latter of whom find a size r ≈ 6.5 pc.

To the southwest of the main cluster IC 1590 is a dark lane known as NGC 281-W. Another dark area just to the east of the nebulosity is designated NGC 281-E. These are the locations of molecular clouds at the same distance as the large nebula, and presumably situated, at least in part, on its near side. APOD provides an optical picture ⁹ that may clarify the geometry for some readers. CO maps of the overall region can be found in Elmegreen & Lada (1978) and Lee & Jung (2003), who provide evidence for cloud compression coming from the direction of the stellar cluster and H ii region.

The NGC 281-W molecular core has a bright submillimeter and far-infrared source, readily apparent in Herschel/SPIRE (Griffin et al. 2010) and SCUBA-2 (Holland et al. 2006) maps. This dust clump harbors a collection of mid-infrared point sources to its east and southeast, as seen in WISE (Wright et al. 2010) images.

The portion of this highly embedded cluster that is visible in the near-infrared was discovered and studied by Carpenter et al. (1993). These authors targeted the region with early near-infrared imaging capabilities based on the position of IRAS 00494+5617. They found a cluster size r = 0.74 pc (corrected here for differences in the distance assumption), with northeastern and southwestern portions separated by a ridge of high extinction. The "typical" extinction toward the embedded sources was estimated at A_V ≈ 10 mag, with peak A_V ≈ 45 mag. Carpenter et al. (1993) also derived a mass of 340 M_⊙ (corrected here for differences in the distance assumption), for the associated CS core. ¹⁰ The same area of NGC 281-W was specifically targeted in a later similar study by Megeath & Wilson (1997) using more sensitive radio and infrared observations.

The object of interest here, PGIR 20dci, is located within the region of these previous surveys, about 30'' west of the embedded cluster center. Figure 1 shows a portion of the field imaged by Megeath & Wilson (1997), with the position of PGIR 20dci designated. Although faintly visible at $K^{\prime}$ band in this archival image, PGIR 20dci has not been previously cataloged, studied, or characterized.

Zoom In Zoom Out Reset image size

In this section, we describe the initial photometric alert that drew our attention to PGIR 20dci. We then detail the optical, near-infrared, and mid-infrared photometric forensics work leading to establishment of the outburst profile over time.

3.1. Gattini 1.2 μm (J-band) Light Curve

The Palomar Gattini InfraRed survey (PGIR; Moore & Kasliwal 2019; De et al. 2020a). observes the sky every two nights in the near-infrared J band, to a median depth of J = 15.7 mag (AB). The faint limit is much brighter than this in crowded or confused regions, such as near the Galactic plane and in high surface brightness regions, such as NGC 281-W.

On 2019 August 31, a "hostless" source was first detected in PGIR as a positive image subtraction and flagged (by coauthor K.D.) for photometric and spectroscopic follow-up. The initial source position required some refinement owing to the 8'' pixel size of PGIR. There are no counterparts in optical (e.g., PanSTARRS; Flewelling et al. 2020) or near-infrared (Two Micron All Sky Survey (2MASS); Skrutskie et al. 2006) survey catalogs, though a hard stretch of the 2MASS images does reveal a faint detection at K_s and perhaps H. A nearby WISE catalog source at 00:52:20.41, +56:34:03.3 was determined to be the source that had brightened. The brightening is also recorded in NEOWISE (Mainzer et al. 2014) data. The consequent increase in signal-to-noise ratio (S/N) likely yields better astrometry and hence a more accurate true source position of 00:52:20.21, +56:34:03.9, which we adopt. This position is 13 away from the position reported for the fainter WISE source, and 4''–5'' from the PGIR position.

Figure 2 shows the light curve produced from the PGIR forced photometry pipeline. Point-spread function (PSF) fit photometry is performed, with noise uncertainty at the location of the transient estimated, as described in De et al. (2020b), from difference images produced using the ZOGY algorithm (Zackay et al. 2016).

Zoom In Zoom Out Reset image size

Since its discovery at J = 14.97 mag, PGIR 20dci has slowly increased in brightness. The brightening trend observed during the 2019–2020 season (Figure 2) can be fit with a linear slope of −0.116 ± 0.011 mag month⁻¹. During the latter part of 2020, the light curve appears to have plateaued, and a linear fit is consistent with zero slope.

3.2. Prediscovery Photometric Information

3.3. WISE+NEOWISE Light Curve

In this section, we present mid-infrared data that support the PGIR detection of a significant brightening of the source.

We collected photometry from the WISE Multi-Epoch and NEOWISE (Cutri et al. 2015) mission archives via IRSA, ¹⁴ rejecting measurements with qi_fact = 0 or qual_frame = 0. We checked the WISE/NEOWISE photometry flags, ¹⁵ finding no entries in the w[12]cc_map_str (contamination and confusion) column, and only nonevent entries in the na (deblending) column. The nb (multiple PSF) column indicates that a single profile is used in >80% of the measurements; only <20% of the measurements (all toward the fainter end) require a second PSF, perhaps to account for the faint star just to the north of PGIR 20dci (see the Spitzer image panel in Figure 3).

Zoom In Zoom Out Reset image size

For the WISE Multi-Epoch catalog and the subsequent, ongoing NEOWISE monitoring data, we took the median of all profile fitting magnitude measurements within each spacecraft "visit" covering the position, and we calculated the error as the dispersion among these magnitudes. The median time span per WISE or NEOWISE visit is about 1.25 days. Results are provided in Table 1.

As illustrated in Figure 3, the mid-infrared light curve of PGIR 20dci experienced an exponential rise phase over the past few years, which appears to be plateauing during 2020. Considering only the three NEOWISE measurements at MJD = 58491.8, 58698.2, and 58858.1, before the flattening, the rise rate was −0.121 ± 0.015 mag month⁻¹ or −1.40 mag yr⁻¹ during 2019–2020. We note that this is identical, within the errors, to the slope derived for the J-band light curve (−0.116 ± 0.011 mag month⁻¹; see discussion above regarding Figure 2). Before this steep recent rise, there was a much shallower but steady upward trend in the WISE/NEOWISE light curve between 2010 and 2019, with a rise slope of only −0.01 mag month⁻¹ or −0.13 mag yr⁻¹.

3.4. Overall Mid-infrared Light and Color Evolution

In this section, we describe the imaging and spectroscopy that we collected on PGIR 20dci.

4.1. Imaging

4.1.1. Data Acquisition

PGIR 20dci was flagged for follow-up with the Palomar 60'' telescope and SEDM (Spectral Energy Distribution Machine; Blagorodnova et al. 2018) in its imaging mode. A sequence in the gri filters was acquired on 2020 August 14 (UT) and reduced using the photometric reduction pipeline.

On 2020 August 28, we obtained near-infrared imaging of PGIR 20dci using the Wide Field Infrared Camera (WIRC; Wilson et al. 2003) on the Palomar 200-inch telescope. The data were acquired as a series of dithered exposures in J, H, and K_s bands for a total exposure time of 495, 330, and 300 s, respectively. The data were reduced, stacked, and photometrically calibrated using the pipeline described in De et al. (2020a).

4.1.2. Findings

No counterpart to PGIR 20dci was detected in g and r bands, with an approximate upper limit of >21 mag. However, we detect a faint extended source in i band, coincident with the bright nebula detected in the near-infrared image. The final K_s -band image was shown in Figure 1. A close-up on PGIR 20dci appears in Figure 4, which highlights the extended cometary structure of the source in the near-infrared. Figure 5 shows the JHK color composite image.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Taken together, the optical and infrared imaging suggests an extremely red source, as well as a bright, extended, scattered light component associated with the brightening of PGIR 20dci. The angular size of the extended structure corresponds to a physical size of ∼14,000 au.

In order to encompass the non-point-source, extended nature of PGIR 20dci, we measured photometry from the Palomar/WIRC images using a fairly large aperture radius of 9 pixels (225), whereas a 5-pixel (125) radius would suffice for point sources. Using the smaller radius would decrease the brightness of PGIR 20dci by about 0.1 mag. We calibrated by matching to 110 sources in common between our final images and the 2MASS catalog. We find J = 14.83, H = 13.08, and K_s = 11.77 mag in 2020 September. The near-infrared colors of PGIR 20dci in outburst are thus J − H = 1.75 mag and H − K_s = 1.31 mag. This location in a J − H versus H − K diagram is consistent with that of a young stellar object that is highly reddened.

4.2. Spectroscopy

4.2.1. Data Acquisition

On 2020 August 13, PGIR 20dci was observed with Keck II and Near Infrared Echellette Spectrometer (NIRES), which is part of the family of similar instruments described in Wilson et al. (2004). The detector records a prism-dispersed simultaneous YJHK spectrum at resolution R ≈ 2700. Four spectra were acquired using a 6'' ABBA nodding pattern, with integration time of 90 s at each position. The data were reduced using a modified version of spextool (Cushing et al. 2004) and flux-calibrated with the telluric A0V standard star HIP 6002 using the xtellcor code (Vacca et al. 2003). Figure 6 shows the final extracted and combined spectrum. Due to the very red nature of the source, there is little signal at the shorter wavelengths, and thus we show only the JHK spectral region.

Zoom In Zoom Out Reset image size

On 2020 September 4, a second spectrum was obtained at R ≈ 2700 with the Palomar 200'' and the TripleSpec (Herter et al. 2008) near-infrared spectrograph. Although with a lower S/N, and not shown as a result, the spectrum is broadly the same in its shape and presence of strong spectral features as the Keck/NIRES spectrum from a month earlier. However, the flux level is brighter. Considering Figure 2, the flux at this point in time could have been several tenths of a magnitude brighter, though the error on this estimate is large. A more likely explanation is that the TripleSpec slit (10) is wider than the NIRES slit (055), and thus the spectrum would include more of the bright extended structure of the source (see Figure 4). Normalizing the two at K band, the K-band slope is identical to that shown in Figure 6, but there is a sharper break just before the CO region, resulting in even deeper CO absorption. The H-band region is then slightly brighter than the spectrum shown in Figure 6. This could mean that the source became slightly bluer while temporarily brighter at this epoch. However, we ultimately attribute the difference to the possibility that the two spectra may have included different parts of the extended nebula.

4.2.2. Findings

Focusing on the Keck/NIRES spectrum (Figure 6), PGIR 20dci is highly reddened, consistent with the extremely red infrared colors measured by WIRC and WISE. The comparison objects ¹⁸ in Figure 6 are less reddened than PGIR 20dci. We can estimate the relative reddening, and thus the extinction, by applying a standard extinction law to find the best spectral match to the V1057 spectrum. We find a differential extinction ΔA_V = 16 mag, which can be added to a baseline A_V = 4 mag for V1057 (Connelley & Reipurth 2018). The total extinction toward PGIR 20dci is thus estimated to be A_V ≈ 20 mag.

Spectroscopically, the most salient feature of PGIR 20dci is the strong CO absorption. The source has both the Δν = 2 CO bands in the K band and the Δν = 3 bands CO in the H band. There is also strong H₂O in J band and H band, and probably in K band as well, and likely OH in H band. Less obvious is the expected TiO and VO absorption, though it may be weakly present. PGIR 20dci shares with Gaia 17bpi and V1057 Cyg in having clear J-band H₂O absorption and H- and K-band CO absorption. The H₂O is weaker in PGIR 20dci, while the CO is stronger. We note that the CO in V1057 Cyg is atypically weak for FU Ori stars (Connelley & Reipurth 2018).

In terms of atomic lines, weak Na i, Mg i, Al i, Si i, and Ca i absorptions are all present. There is also weak H i Paβ absorption at 1.28 μm. The metal line strengths are similar to those seen in the comparison object Gaia 17bpi and somewhat stronger than in V1057 Cyg. PGIR 20dci and Gaia 17bpi also have comparable Paβ, which is likely blueshifted and formed in a wind.

Notably, the He i λ10830 line in PGIR 20dci shows a clear P Cygni profile (see inset panel within Figure 6). This is evidence of a strong wind or outflow, with blueshifted absorption in He i λ10830 frequently seen in FU Ori stars (Connelley & Reipurth 2018). We estimate a maximum velocity around −450 km s⁻¹.

In the nomenclature of Connelley & Reipurth (2018), PGIR 20dci appears to be a bona fide FU Ori star. In addition to having had its eruption observed, it satisfies all of the spectroscopic criteria: CO absorption, H₂O absorption, VO or TiO absorption (weakly), Paβ absorption, lack of emission lines, weak metal absorption, and He i λ10830 absorption. The hydrogen and helium lines likely have some outflow aspect to their absorption. The metal lines originate in the source photosphere, however, and combined with the molecular line absorption, they indicate a mixed-temperature spectrum.

The dramatic brightening behavior of PGIR 20dci is detected most obviously in the mid-infrared (Figure 3). The combined Spitzer, WISE, and NEOWISE data provide a long-duration light curve. As the source has brightened, it has become detectable in the near-infrared as well (Figure 2). In the PGIR data, which are higher cadence, the brightness of PGIR 20dci is currently near the survey limit owing to the high background in this field, and the light curve is thus somewhat noisy, though well sampled. Over only the past few months, PGIR 20dci has appeared near the faint limit in the ZTF alert stream, sampling the red optical. Given the evidence for a current (second) plateau phase, the source may not become amenable to study in the optical, however.

A two-step rise is indicated by the mid-infrared light curve, each step being preceded by a shallow rise phase. The earliest available data for PGIR 20dci show a shallow brightening between 2004 and 2009. Then, PGIR 20dci exhibited a rapid (−0.25 mag month⁻¹) blue brightening by 1.8 mag in W1 and 0.8 mag in W2 between 2009 September 6 and 2010 January 26. This was followed by a shallower-slope (−0.01 mag month⁻¹), long-term colorless brightening over 2010–2018, amounting to only another 0.5 mag in each of W1 and W2. The slope of the brightening increased during the latter part of 2018 and switched in 2019 to a much steeper rise rate (−0.12 mag month⁻¹). At this point, the source became detectable by PGIR in the near-infrared. PGIR (1.2 μm) and NEOWISE (3.4 and 4.6 μm) both record the same rise slope in this time frame, and both show a transition to a second plateau during 2020. The total brightening of PGIR 20dci over the past 16–25 yr is ∼5−5.5 mag.

Putting these light-curve results into context, the recent rise slope compares well to the typical rise rates that can be derived for the optical light curves of historically documented FU Ori outbursts. These have a wide range but are most frequently about −0.05 to −0.2 mag month⁻¹ (L. A. Hillenbrand 2021, in preparation). Regarding the two-step rise in the mid-infrared light curve, a similar phenomenon has been recorded on one other source: the confirmed FU Ori source Gaia 17bpi (Hillenbrand et al. 2018). However, in that object, the total duration of the first plateau phase was shorter, only 1–2 yr, instead of the 5–10 yr measured here for PGIR 20dci.

We can estimate a source luminosity near the outburst peak based on the brightness level of the apparent plateau in the light curve (J ≈ 14.6 mag from PGIR and r ≈ 20.5 mag from ZTF) and the adopted distance d = 2.81 kpc. The bolometric correction is of course unknown, but adopting values appropriate for a mid-F-type star leads to a luminosity estimate of 10–12 L_⊙, with perhaps a factor of 3 uncertainty if the bolometric correction is ±1 mag. This is on the far low side of FU Ori outburst peak luminosities but notably does not include the long wavelength contribution, which is currently unmeasured. Just Gaia 17bpi, the only other source known to have exhibited a two-step mid-infrared brightening, like PGIR 20dci has, and HOPS 383, an embedded young star outburster whose status as an FU Ori–type outburst is unclear, have luminosities this low.

Concluding our discussion, we note that the best-studied FU Ori stars are the optically visible portion of the population, especially those bright enough for high-dispersion spectroscopy. However, a sizable fraction of the known FU Ori stars, about one-third, are seen only at infrared wavelengths owing to high extinction (e.g., the A_V = 20–50 mag population in Connelley & Reipurth 2018). Indeed, theoretical predictions are that earlier-stage young stellar objects experience a higher rate of FU Ori–type outbursts (Bae et al. 2014). Most of the known optically invisible group have been designated as FU Ori objects through source brightening at K band that was noticed only post facto, rather than while it was actually happening, along with spectroscopic confirmation. Embedded FU Ori sources thus generally do not have quiescence-to-rise light curves available.

Because it meets all applicable criteria for the FU Ori designation, notably the spectral absorption signatures, a lack of emission lines other than P Cygni structure, and the total light-curve amplitude, PGIR 20dci can be added to the list of optically invisible FU Ori candidates. Unlike the others, however, this source has had its photometric rise captured through wide-field continuous photometric monitoring programs.

We have discovered an outbursting young stellar object first detected at near-infrared wavelengths, but also seen in long-duration mid-infrared time series data. PGIR 20dci is associated with the NGC 281-W molecular core, a known area of recent and ongoing dense star formation activity. Its salient features are as follows:

• 1.  
  A Class I spectral energy distribution.
• 2.  
  Foreground interstellar, molecular core, and circumstellar extinction amounting to A_V ≈ 20 mag.
• 3.  
  A photometric rise that was detected in the PGIR time-domain survey and emphasized by comparing the recent J-band photometry with the previous lack of a J-band detection at the same position, implying a total near-infrared brightness increase of >5 mag.
• 4.  
  Extended nebulosity in our new K_s -band imaging, which was not present in 1990s images at the same wavelength, with a cometary-type structure ∼14,000 au in size.
• 5.  
  A measured 10–15 yr photometric brightening trend exhibited in 3–5 μm Spitzer + WISE + NEOWISE data, accumulating a total mid-infrared brightness increase of >5 mag.
• 6.  
  Steepened mid-infrared brightening during 2019 that was coincident with source detection at J band.
• 7.  
  Consistent rise rates during 2018–2020 in the mid-infrared NEOWISE light curve and during 2019–2020 in the near-infrared PGIR light curve, amounting to −0.12 mag month^–1.
• 8.  
  A flattening of the light curve from its rise phase to a plateau in 2020. The peak luminosity is estimated at a modest 10–12 L_⊙, though this is likely a lower limit.
• 9.  
  A near-infrared absorption-line spectrum showing hydrogen line, metal line, and molecular band absorption, indicative of a "mixed-temperature" spectrum.
• 10.  
  Outflow as evidenced by an He i λ10830 profile, with blueshifted absorption seen to about −450 km s⁻¹.

We conclude that PGIR 20dci is a bona fide FU Ori star. The object warrants further study, especially at high spectral dispersion, as its outburst continues to develop.

Palomar Gattini-IR (PGIR) is generously funded by Caltech, Australian National University, the Mt. Cuba Foundation, the Heising Simons Foundation, and the Bi- national Science Foundation. PGIR is a collaborative project among Caltech, Australian National University, University of New South Wales, Columbia University, and the Weizmann Institute of Science. M.M.K. acknowledges generous support from the David and Lucille Packard Foundation.

Facilities: Palomar Gattini IR - , PO:1.2 m:ZTF - , PO:1.5 m:SEDM - , Hale:WIRC, Hale:TSPEC - , Keck/NIRES - , Keck/DEIMOS - , Spitzer - , WISE - , NEOWISE - , IRSA. -