Gaia 19ajj: A Young Star Brightening Due to Enhanced Accretion and Reduced Extinction

Gaia 19ajj is located within an optically opaque small cloud, positioned about 15 below the southern galactic plane. Figure 1 shows the large-scale region. The sightline passes the northern part of the Vela Molecular Ridge, broadly located toward the Gum Nebula. The best known objects in this general area are the CG 30/31 globules containing HH 120, located 15' to the west (Reipurth 1983). About a degree to the northeast is the prominent H ii region RCW 19 = Gum 10 (Gum 1955; Rodgers et al. 1960).

Zoom In Zoom Out Reset image size

Between these two features is a small cloud, Dobashi 5180 (Dobashi 2011), in which Gaia 19ajj is located. Other very nearby cataloged structures include the dark clouds DC 253.6-1.3 (Hartley et al. 1986) and G253.56-1.32 (Dutra & Bica 2002), and the Planck-identified dust clump and dense core PLCKECC G253.49-01.26 (Planck Collaboration et al. 2011) to the east. The Gaia 19ajj position is just southeast of IRAS 08088-3554 which is identified as a compact H ii region.

The immediate region of Gaia 19ajj features many diffuse objects in optical images from the National Optical Astronomy Observatory (NOAO) DECam Plane Survey (DECaPS; Schlafly et al. 2018), and there is evidence of a grouping of bright red stars, including Gaia 19ajj, in the mid-infrared WISE images. The local cloud is clearly a recently active star-forming region. The most well-studied young star in the vicinity is HBC 556, a few arcmin to the west. Pettersson (1987a, 1987b) identified this source as PHα21, of spectral type M4, in a broad-area search for Hα emission stars in the region between the cometary globules and RCW 19. In order to gauge the level of star formation activity in this region, we looked for the presence of additional Hα emission sources in the vicinity of Gaia 19ajj by examining an Hα image obtained as part of the AAO SuperCOSMOS survey (Parker et al. 2005). We identify four such sources in Figure 1. A color image from DECam (available via the ALADIN service) shows even more detail of these and additional, fainter nebulae in the vicinity.

The Gaia DR2 (Gaia Collaboration et al. 2018) distance to Gaia 19ajj, derived by simply inverting the measured parallax (1.200 ± 0.201 mas), is ${832}_{-120}^{+170}$ pc. This can be compared to previous estimates of distance in this region of sky. First, Westerlund (1963) discovered an OB association, which he called Puppis OB III, apparently related to the RCW 19 H ii region to the east of Gaia 19ajj, and estimated its distance at around 1700 pc. Considering the more proximately located HBC 556, Pettersson (1987a) assumed that this star is located within the Gum Nebula, at the same approximate distance of the cometary globules to the west of Gaia 19ajj, estimated at around 450 pc. We concur that from its brightness and spectral type, HBC 556 cannot be a distant object. Finally, a kinematic distance of 3.3 kpc was suggested by Bhatt et al. (1998) for the associated small cloud itself, but this value is both highly uncertain and improbable given the other evidence. We adopt the 832 pc distance from Gaia DR2. This implies that Gaia 19ajj, and thus presumably the Dobashi 5180 small cloud, are located behind the Gum Nebula.

The spectral energy distribution (SED) that can be assembled for Gaia 19ajj (Figure 2; see also the Vizier version⁷ ) rises from 0.5 to 4 μm, then turns over with flat or only slightly declining energy out to 24 μm. The SED is comprised of photometry taken at different epochs—when the source was in various photometric states. Catalog data originates from the The Naval Observatory Merged Astrometric Dataset (NOMAD; Zacharias et al. 2005), USNO-B (Monet et al. 2003), VPHaS DR2 (Drew et al. 2014), Gaia DR2 (Gaia Collaboration et al. 2018), Two Micron All Sky Survey (2MASS; Cutri et al. 2003; Skrutskie et al. 2006), Deep Near Infrared Survey of the southern sky (DENIS),⁸ WISE (Wright et al. 2010), Midcourse Space Experiment (MSX; Price et al. 2001), and AKARI (Ishihara et al. 2010) catalogs. The $G,{BP},{RP}$ photometry from Gaia DR2 is an order of magnitude fainter than the somewhat earlier VPHaS optical photometry and much fainter than the decades earlier USNO-B value (see Figure 3).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Regardless of the demonstrated short-wavelength photometric variability, Gaia 19ajj is clearly a "Class I" type young stellar object with a flat mid-infrared SED.⁹ This SED category is associated with disk+envelope circumstellar geometry, suggesting a significant amount of high latitude material above the disk that is distributed in a more spherical-like geometry, with ongoing infall and rapid accretion onto the central stellar object. We do not attempt to model the SED given the demonstrated variability and the geometric complications for the radiative transfer.

The SED also implies that at least some of the short-wavelength light may be scattered rather than direct. This is consistent with the fact that the fainter-state photometry represented in the Gaia DR2 data does not demand significantly higher reddening relative to the much brighter state of the VPHaS measurements. As discussed below, the photometric changes in Gaia 19ajj are likely not explained by extinction changes alone.

There are no previous publications studying this specific star (2MASS 08104579-3604310), though it does appear based on its WISE colors in the large catalog of Marton et al. (2016) identifying candidate young stellar objects.

Gaia 19ajj is an apparent Hα emitter, even in previously fainter states. The VPHaS catalog (Drew et al. 2014) provides a brightness of r = 16.86 mag and a moderate r − i color of 1.13 ± 0.02 mag, along with an emission index of $r-{\rm{H}}\alpha =0.89\pm 0.02$ mag. According to the models in Figure 2 of Barentsen et al. (2013), the source would be an obvious Hα emitter, with an equivalent width of several tens of Angstroms.

Given the SED, as well as the large-amplitude photometric variability described below, it is unwise to attempt a luminosity calculation. However, the observed optical colors suggest a spectral type earlier than ∼M3, assuming there is also some color contribution from reddening. Adopting the ≈830 pc distance, the absolute magnitude in an average part of the lightcurve (see below) is M_r = 7.26 − A_r, where A_r is the unknown value of the extinction in the r-band. A typical luminosity around 1 L_☉ for a young pre-main-sequence star would suggest A_r ≈ 3 mag.

Figure 3 shows the Gaia G-band lightcurve¹⁰ with the last update for our analysis occurring on 2019 February 28. No error estimates are provided but we adopt 0.05 mag uncertainty. We include an VPHaS (Drew et al. 2014) r-band measurement that has been approximately converted to the Gaia photometric system (using the procedure described in Hillenbrand et al. 2018). Also shown is the V-band and g-band data available from ASAS (Shappee et al. 2014). These data have higher cadence than the Gaia data, and indicate quasiperiodic oscillations during the 2018–2019 rise, with a periodogram peak around 24 days. We investigated the availability of Transiting Exoplanet Survey Satellite (TESS) data for this source, which was observed from 2019 January 7 through February 28, but the TESS pixels are too big given the field density for a viable lightcurve to be derived at the current level of TESS data processing.

The total brightening of Gaia 19ajj over the past three years exceeds 5.5 mag in the Gaia G-band. While impressive, and possibly suggestive of an outbursting source such as an FU Ori star, Gaia 19ajj was also bright in early 2012, conceivably as bright as it is now. The subsequent large-amplitude fade and more recent re-brightening at both infrared and optical wavelengths suggests that the photometric excursions may have a complex interpretation, such as being due in part to accretion variations and in part to extinction variations. A long-duration color time series would help in the assessment. Unfortunately, there is little optical color information available at present. By matching the time series between Gaia and ASAS, we derive only a crude g − G color estimate in the range of 1.2–1.8 mag and no meaningful color variability information.

The historical record of photographic plate measurements of Gaia 19ajj supports repeated large-amplitude brightening and fading. The USNO-A2 (Monet 1998) records R = 17.25 mag for a mean epoch between blue and red observations of 1978.5, while the USNO-B1 (Monet et al. 2003) gives R = 14.66 mag for mean epoch 1982.3. In other words, the source has exhibited approximately 3 mag variability on few year timescales in the past, similar to what has occurred over the past 4–7 yr.

Variability is also evident in mid-infrared lightcurves. Gaia 19ajj is well detected in data from both the primary WISE mission (Wright et al. 2010) and the NEOWISE Reactivation mission (Mainzer et al. 2014).

The 3.4 and 4.6 μm (W1 and W2) profile-fit photometry for Gaia 19ajj was drawn from the WISE All-Sky, 3-Band Cryo, Post-Cryo, and NEOWISE Reactivation Single-exposure Source Databases. Measurements were used only from exposures for which the image quality parameter qual_frame was greater than zero, and on which the source lies more than 50 pixels from the image edge. Gaia 19ajj is formally saturated in all WISE/NEOWISE images, requiring additional attention in order to present quality photometry. Specifically, the profile-fit photometry of saturated sources exhibits a flux-dependent bias of varying degrees. Therefore, we derived and applied corrections, now published,¹¹ to the source database magnitudes in order to compensate for the photometric bias.

The median values of the corrected W1 and W2 magnitudes for each observation epoch were computed. The uncertainties associated with each median magnitude are the standard deviation of the population of the corrected magnitudes within each epoch. These values are typically slightly smaller than the median of the uncertainties (error) of the individual corrected magnitudes that include the contribution of the systematic uncertainties of the photometric bias corrections. WISE/NEOWISE photometry is provided in Table 1.

As illustrated in Figure 3, the WISE data points from 2010 are relatively bright, while NEOWISE measured fading between 2014 May and November, then gradual brightening from about MJD ≈ 57,000 to the present. There is superposed variability that generally traces the pattern of the optical fluctuations, but at lower amplitude. The cumulative brightening since 2014 November is about 1.2 mag in W1 (3.6 μm) compared to >5 mag in G (0.67 μm).

Figure 4 shows the WISE/NEOWISE color curve which indicates that the source has been getting gradually bluer since about MJD ≈ 57,000, while it has become brighter (Figure 3). Figure 4 also shows that there is a direct and tight correlation between color and magnitude. However, a least-squares fit indicates that the magnitude–color slope is relatively shallow at W1 = 2.39 × (W1 − W2) + 4.29 with rms = 0.10 mag. This can be compared to the slope of 4.31 expected for interstellar extinction (Indebetouw et al. 2005, adopting the Spitzer extinction law as approximately correct for the WISE bands). If interpreted as extinction, the total color change would correspond to A_V = Δ(W1 − W2)/(0.56 − 0.43) × 8.8 = 67.7 × Δ(W1 − W2) ≈ 35 mag. However, there is more color change than expected given the magnitude change, which suggests that at least some of the blueing behavior is intrinsic to the source and not just extinction clearing.

Zoom In Zoom Out Reset image size

This shallow color slope is also apparent in the near-infrared. As discussed below, comparing the changes in J − H versus H − K between 2MASS and new Infrared Telescope Facility (IRTF) photometry shows larger H − K change relative to J − H than would be expected under standard interstellar reddening assumptions. As noted above, the SED also shows an inconsistency between the large change in optical brightness relative to only a small change in optical color.

Both the Gaia lightcurve and the NEOWISE lightcurve exhibited quasiperiodic behavior over the few years prior to the recent steep optical rise. Then, starting in early 2018 and continuing into 2019, the slope changed and Gaia 19ajj brightened by another several magnitudes (∼3.5 in the optical and ∼1 in the infrared). The optical slope was −0.31 mag/month over about 10 months.

The optical and mid-infrared lightcurve evidence for wavelength-dependent, long-term photometric brightening that is punctuated by brief fading events suggests that extinction variability may be an important consideration in interpreting the behavior of Gaia 19ajj. However, extinction variations are not the entire story. Accretion variability is driving some of the brightening and blueing, which also has the effect of reducing the amount of dust, and thereby diminishing the line-of-sight extinction. This is the same interpretation as is invoked for several better-studied sources with well-characterized large-amplitude and long-timescale lightcurves, like V2492 Cyg and PV Cep (see the discussion in Section 6).

We obtained near-infrared spectra using the IRTF and SpeX (Rayner et al. 2003) on 2019 February 3 (UT) in both short wavelength cross-dispersed (SXD; 0.7–2.4 μm) and long wavelength cross-dispersed (LXD; 1.7–4.2 μm) modes. The resolution with the 05 slit was R ≈ 1200. For the SXD data the realized signal-to-noise ratio (S/N) per exposure is ∼80 in the Y-band and ∼300 in the K-band (with 10 exposures taken for 30 minutes of total exposure time). For the LXD data the S/N is ∼50 per exposure in L' (with 16 exposures taken for 8 minutes of total exposure time). The K-band seeing was about 05 at zenith but closer to 1'' for observations taken over 1.78 to 1.95 airmasses at the target position. A second SXD spectrum was obtained on 2019 March 26 (UT) under somewhat worse seeing conditions. All data were reduced with SpeXtool (Cushing et al. 2004). In addition to the spectroscopy, a small set of H₂ and broadband K images were obtained on the March 26 observing date, and show spatially extended H₂ emanating from the southeast (despite the poor seeing of ∼19), extending to about 7''.

We obtained an optical echelle spectrum between ∼3400–7900 Å at a resolution of R ≈ 60,000 using the Keck I telescope and High Resolution Echelle Spectrometer (HIRES; Vogt et al. 1994) on 2019 February 14 (UT). Data acquisition used the standard operating procedures of the California Planet Search described in Howard et al. (2010). Under temporarily clear skies and ∼20 seeing, a 1500 s exposure resulted in a spectrum with S/N = 15 at 5500 Å. We aligned the Li i absorption feature to determine a source radial velocity of +35.8 km s⁻¹.

Finally, we obtained an infrared echelle spectrum in the 1 μm Y-band region at R ≈ 37,000 with the Keck II telescope and the recently upgraded (Martin et al. 2018) Near-Infrared Spectrograph (NIRSpec; McLean et al. 1998) on 2019 April 8 (UT). Two rounds of A–B–B–A position nods were taken with individual exposure times of 150 s per nod position. The seeing was 05–06 overhead, though the target was observed at 1.8 airmasses. These data were processed using the REDSPEC package¹² and the resulting S/N = 60 at 1.08 μm.

Figures 5–11 illustrate some details of our spectra and are discussed in the subsections below, with the main features as follows. The Gaia 19ajj optical and infrared spectra exhibit many prominent lines that are indicative of outflowing material. There is evidence for both a neutral stellar/disk wind absorbing continuum radiation, and also shocked gas indicative of an outflowing jet. Weak metallic emission lines appear at the rest velocity.

Zoom In Zoom Out Reset image size

In the optical, notably absent is strong photospheric absorption, with a few weak metallic lines identified, as discussed below. Several absorption features can be attributed to diffuse interstellar bands. In the infrared, the SXD observations in the YJHK bands indicate a red continuum with broad absorption due to molecular species. The LXD observations in the L-band exhibit a flat/blue continuum shape.

5.2.1. Wind Lines

Gaia 19ajj exhibits classic P Cygni structure in the Balmer Hα feature, with redshifted emission peaking around +25 km s⁻¹ and blueshifted absorption extending to −450 km s⁻¹ (Figure 6). Similar structure is seen more weakly in the infrared in the Paβ, Paγ, and Paδ lines, albeit from much lower spectral resolution (Figure 5). At Brγ there is simple emission, and Brδ and Br10 are in absorption, with very weak absorption or no sign of higher Br lines.

Zoom In Zoom Out Reset image size

There is also a P Cygni signature in the Ca ii infrared triplet lines around 8500 Å (Figure 5), which is likely present as well in the Ca ii K line at 3933 Å (though masked by low S/N; Figure 6). Other lines with wind signatures in the form of purely blueshifted absorption include: Fe ii 5018, 5169, and 5316, the Mg i b triplet, the Na i D doublet, the K i 7665/7699 doublet, the O i 7774 triplet (Figure 6), and the higher excitation He i 10830 triplet (Figures 5 and 11).

The hot He i line is very strong, reaching 50% of the continuum level in the earlier SpeX spectrum and all the way to 5% of the continuum in the later NIRSPEC spectrum, with the terminal velocity around −450 km s⁻¹. This assumes that the rest velocity is at a heliocentric velocity of +35.8 km s⁻¹ which was the optical correction based on the Li i line, and indeed matches the location of the photospheric Mg ii + Sr ii line in the 1 μm spectrum. There is no signature of a feature in either absorption or emission at He i 5876 Å, which is the next transition higher from the 10830 Å upper level, and is optically thin. The wind models of Kwan et al. (2007) which treat the He i 10830 line formation as a purely scattering process from the continuum photons are consistent with our observations. More specifically, the breadth and depth of the Gaia 19ajj line profile (Figure 11) seem to match the genre of Kwan et al. (2007) models that adopt a stellar wind geometry rather than an inner disk wind geometry. Empirically, the He i profile of Gaia 19ajj most resembles that of the strong outflow source SVS 13, as illustrated in Edwards et al. (2006).

5.2.2. Jet Lines

Shocked outflowing gas is evidenced by blueshifted and broad emission in the optical doublets of [O i], [S ii], [Ca ii], [Fe ii], and [Ni ii]. As illustrated in Figure 7, all of these lines are double-peaked with the [S ii] triply peaked. In most species the higher velocity component is stronger, but in [Ca ii] the lower velocity component is stronger. Uniquely for the [Ca ii] doublet lines, the low-velocity peak is at zero velocity relative to the star.

Zoom In Zoom Out Reset image size

In the infrared (Figure 5), jet signatures manifest in several lines of [Fe ii], along with shocked H₂ emission in the lowest energy lines. The O i line at 1.13 μm is somewhat broad, and has been associated with jets in other young sources.

The [Ca ii], [Ni ii], [Fe ii], and [S ii] lines exhibited in the optical and infrared spectra of Gaia 19ajj have intermediate ionization potential, 6.1–10.4 eV. There is no evidence of the often-seen lines of [O ii] or [N ii] which have higher ionization (13.6–14.5 eV), suggesting moderately low temperature. The various ratios among same species lines appear consistent with intermediate density ($\gt {10}^{4}\mbox{--}6\times {10}^{5}$ cm⁻³)—except for the [Ca ii].

The [Ca ii] lines are high density (Ferland & Persson 1989; Hartigan et al. 2004; Nisini et al. 2005) and are unusual in T Tauri stars, even those showing the forbidden lines discussed above. They have been observed, however, in a number of unusual large-amplitude photometric variables such as RW Aur, V2492 Cyg (PTF 10nvg), and PTF 14jg (Hillenbrand et al. 2019), as well as in the extreme emission-line young star V1331 Cyg. As mentioned above, the [Ca ii] is the only forbidden species in Gaia 19ajj having a zero-velocity component (Figure 7). The ratio of the 7291 Å to the 7324 Å is approximately 1.2:1, slightly lower than the expected 1.5:1 discussed in great detail by Hartigan et al. (2004). The [Ca ii] lines are ground state transitions with their upper level populated either by collisions upward, or by downward transitions via the permitted Ca ii triplet lines. If the [Ca ii] occurs by fluorescence rather than collisions, this may explain both the different morphology relative to the other forbidden lines, and the slightly low doublet ratio.

5.2.3. Narrow Emission and Weak H i Emission Lines

There is narrow rest-velocity emission in Gaia 19ajj in permitted metallic lines such as Fe i, and Ni i in the optical, as illustrated in Figure 8. There is also weak metal emission in the infrared, as illustrated in Figure 5, which shows these same species plus other lines such as Mg i and Si i in J-band and H-band, and Na i in K-band. The line strengths are generally only W_λ ≳ − 1.0 Å. This narrow emission spectrum is similar to that of the pre-peak spectrum of the FU Ori candidate outbursting source PTF 14jg (Hillenbrand et al. 2019), the large-amplitude accretion+extinction source V2492 Cyg (PTF 10nvg; Hillenbrand et al. 2013), as well as that of the photometrically quiet but extreme emission-line star V1331 Cyg.

Zoom In Zoom Out Reset image size

The infrared spectrum of Gaia 19ajj also shows weak emission in the CO bandheads, though perhaps superposed on troughs of absorption (see the discussion below).

In the long wavelength spectrum (Figure 9), there is weak H i emission as evidenced by Brα, Pfγ, and Pfδ. There is another weak line at 3.675 μm that we cannot identify (the wavelength seems incorrect for Hu19 at 3.645 μm and there is no corresponding Hu18 at 3.693 μm). The spectrum bears some resemblance in showing these weak lines to V1647 Ori and Z CMa (Connelley & Reipurth 2018).

Zoom In Zoom Out Reset image size

5.2.4. Interpreting the Strong Emission Lines

The strongest emission in Gaia 19ajj comes from Hα and Ca ii, both associated with the accretion flow in typical young stars when these lines are strong and broad. The Hα line has ${W}_{\lambda }\approx -11$ Å in its redshifted emission component, but as discussed above there is also a strong blueshifted absorption component (W_λ ≈ 5 Å) to the overall P Cygni type profile, weakening the emission strength. The 8542 Å line, usually the strongest of the Ca ii triplet lines, has W_λ = −6.3 Å (integrated flux 4 × 10⁻¹⁴ erg s⁻¹ cm⁻² Å⁻¹), while the 3933 Å K line, usually the stronger of the Ca ii doublet, has W_λ ≈ −18 Å (with large uncertainty due to the low S/N in the continuum in this region). Again, there is likely blueshifted absorption in the Ca ii lines that reduces the emission strengths.

We note with some interest the complete absence of the 3968 Å H line (see Figure 10), which is highly unusual given the generally comparable fluxes of the two components of the doublet (K:H ratio in the range 1–2) in most young accretors (e.g., Herczeg & Hillenbrand 2008; Rigliaco et al. 2012). In active chromospheres the K:H ratio increases as the activity level decreases and also occupies the range 1–2 (Houdebine & Stempels 1997). Suppression of the H line relative to the K line is clear in Gaia 19ajj and seems to defy the atomic physics prediction of a maximum ratio of 2 (in the optically thin case from the g−f values). Despite the doublet ratio oddity, the three lines of the Ca ii triplet, which share upper levels with the Ca ii doublet, have typical ratios for young star accretors, around 0.95:1:0.85 for the 8498:8542:8662 Å lines. The usual interpretation is that the triplet lines are very optically thick (see the expected 0.11:1:0.55 proportions for optically thin lines).

Zoom In Zoom Out Reset image size

An explanation for the peculiar doublet ratio is that there is contaminating P Cygni structure from H just 121 km s⁻¹ redward of the Ca ii H line. However, while Hα is strong in Gaia 19ajj, the upper Balmer lines are weak with only hints of Hγ, Hδ, and H. A complete cancellation of Ca ii H emission having a strength of ∼50%–100% of the Ca ii K line would require a coincidental H blueshifted absorption component of equal or larger strength, but without a redshifted emission component that exceeds the very weak H signal illustrated in Figure 10. We find that Herbig (1989) called attention to the same peculiarity of a lack of Ca ii H despite strong Ca ii K in the pre-outburst spectrum of V1057 Cyg, as well as noting it in FU Ori in the post-outburst state along with a few other unusual young stellar objects. The H line suppression may be a feature of sources with strong winds having a certain temperature and column density or optical depth.

5.2.5. The Photosphere

Gaia 19ajj is compared in Figure 5 to the FU Ori type star V1057 Cyg, which shows rather prominent ${}^{12}\mathrm{CO}$ absorption in the 2.3 μm region (Connelley & Reipurth 2018; though weakening recently). As noted above, Gaia 19ajj shows some signatures of emission near the bandhead regions, which turn into a drop in the continuum toward redder wavelengths due to the blended CO lines. The peculiarity of the CO region aside, there is resemblance of the Gaia 19ajj near-infrared spectrum with that of FU Ori stars in the prominent H₂O absorption feature, which is seen in all of the K-band, H-band, and J-bands (Connelley & Reipurth 2018). We have two SpeX spectra taken two months apart, and they show subtle but significant changes in the overall spectrum. Specifically in this interval, the water vapor absorption bands grew stronger, resulting in a more triangular shape through the H-band. At the same time, the CO emission lines in the K-band became slightly weaker, while the Brγ and the H₂ emission lines became slightly stronger. At longer wavelengths, Figure 9 shows only a flat continuum. There is no broad absorption between 2.8 and 3.3 μm due to water ice, as is seen in many young embedded or Class I type stars (e.g., Connelley & Reipurth 2018).

In terms of narrow atomic absorption, the low-resolution near-infrared spectrum of Gaia 19ajj exhibits little of the absorption due to atomic lines expected in an FU Ori interpretation. Our NIRSPEC spectrum, however, does show a photospheric line (Figure 11) that can be attributed to Sr ii and/or Mg ii, both intermediate ionization species. Wallace et al. (2000) demonstrate that this feature is prominent in F- and G-type supergiants, weaker in giants, and absent in dwarfs. In the optical, aside from the hallmark Li i 6707 Å youth signature (Figure 7), the strongest rest-velocity absorption that we have identified comes from Ca i and Ba ii (Figure 8). Another order contains both Ca i 5857.45 and Ba ii 5853.67. It is unclear why only Ca i absorption is strongly present, and not other lines having similar excitation. One clue may be that these lines have some surface gravity sensitivity for spectral types later than mid-G (Prisinzano et al. 2012). Besides these more obvious lines, only hints of weak absorption from other neutral metals appear in the Gaia 19ajj high resolution spectrum, especially in the 5500–5900 Å region, for example species such as Si, Sc, Mg, and Fe, Ni.

Zoom In Zoom Out Reset image size

Regarding Ba ii, the strongest (though blended) Ba ii feature at 6497 Å unfortunately is between spectral orders in our HIRES spectrum. The illustrated 6142 Å line has W_λ = 0.39 Å, while the 5853 Å line has W_λ = 0.16 Å. These equivalent widths are notably stronger than the Ba ii line strengths of normal dwarf stellar photospheres (of any spectral type). These lines are known to have a positive luminosity effect, and are strong in supergiants (Andrievsky 2006). The measured equivalent widths are about two-thirds of the line strengths seen in FU Ori, V1057 Cyg, and V1515 Cyg. While the Ba ii line is typically seen in FU Ori sources, including in the recently announced objects Gaia 17bpi (Hillenbrand et al. 2018) and PTF 14jg (Hillenbrand et al. 2019); it is not present in other young stars. This is consistent with both the disk atmosphere (Hartmann & Kenyon 1985) and the distended stellar photosphere (Petrov & Herbig 1992) interpretation of the optical spectra of FU Ori objects, each of which are expected to produce low surface gravity spectral signatures.