VVV-WIT-12 and Its Fashionable Nebula: A 4 yr Long-period Young Stellar Object with a Light Echo?

We report the serendipitous discovery of VVV-WIT-12, an unusual variable source that seems to induce variability in its surrounding nebula. The source belongs to the rare objects that we call WITs (short for What Is This?) discovered within the VISTA Variables in the Vía Láctea (VVV) survey. VVV-WIT-12 was discovered during a pilot search for light echoes from distant supernovae in the Milky Way using the near-IR images of the VVV survey. This source has an extremely red spectral energy distribution, consistent with a very reddened (A V ∼ 100 mag) long-period variable star (P ∼ 1525 days). Furthermore, it is enshrouded in a nebula that changes brightness and color with time, apparently in sync with the central source variations. The near-IR light curve and complementary follow-up spectroscopy observations are consistent with a variable young stellar object illuminating its surrounding nebula. In this case the source periodic variation along the cycles produces an unprecedented light echo in the different regions of the nebula.


INTRODUCTION
A few previously unclassified variable sources that we call "WIT" objects -short for What Is This? -have been discovered by the VISTA Variables in the Vía Láctea (VVV) survey (Minniti et al. 2010).These represent a wide variety of extreme or rare astrophysical phenomena, including a very reddened novae or supernova or a protostellar collision (VVV-WIT-01; Lucas et al. 2020), an extragalactic radio source violently variable in the near-IR (VVV-WIT-04; Saito et al. 2019a), another Tabby star or Mamajek object (VVV-WIT-07; Saito et al. 2019b) and a giant star that blinked (VVV-WIT-08 Smith et al. 2021).
In year 2012 we started a pilot search for light echoes in the near-IR images of VVV survey.Light echoes oc-cur when the interstellar dust acts as a mirror reflecting the light of a very bright transient event like a SN (Patat 2005;Patat et al. 2006;Rest et al. 2008Rest et al. , 2011)).We made RGB color images using observations in the K s -band acquired in three different epochs from 2010 to 2012.Very few candidates were detected, probably because the timespan was short.However, one of these candidates showed a nebula with changing colors, around a strongly variable point source that we named VVV-WIT-12 (see Fig. 1).
It is not clear if VVV-WIT-12 produces a light echo or the mere illumination of the gaseous nebula by a strong variable source.Since this object defies classification, we labeled it as one of our WIT objects.This discovery is relevant because the Vera Rubin Observatory (former LSST) will start operations soon, enabling a massive search for light echoes from ancient SNe in the Milky Way (Li et al. 2022).Such searches have the potential to discover and monitor the evolution of other unusual objects like VVV-WIT-12.
The parallax and VVV K s -band light curve of VVV-WIT-12 was obtained from VIRAC2 (Smith et al. 2018(Smith et al. , 2023 in prep.) in prep.).There is a total of 203 well sampled epochs, spanning by 3418 days (∼9.4 years), from April 18 2010 to August 27 2019.The parallax is ω = 2.5 ± 1.5 mas, PM RA = −0.03+ / − 0.33 mas yr −1 , PM DEC = −0.30± 0.33 mas yr −1 .Both the parallax and proper motions are consistent with zero, indicating a distant object.Based on VIRAC2 data, Molnar et al. (2022) classified this object as a long period variable (LPV), reporting a period of P = 292 days.Indeed, our VIRAC2 light curve looks nearly sinusoidal, but with a much longer period of P = 1525 ± 30 days, resembling a LPV (see Fig. 2).
Despite having acquired many observations with the other VVV filters (ZY JH), the central point source is not detected in any of these shorter wavelengths in order to establish if color variations are in synch with the light curve.However, these non-detection allows to estimate the color limits using the limiting magnitudes for the VVV tile b341 Saito et al. (2012).In particular, the extreme (H − K s ) > 5.0 mag suggests very high optical extinction, A V ∼ 100 mag.The mean extinction towards this field, considering a 5 ′ FoV is A V = 23.28 mag (A Ks = 2.69 mag) using the Schlafly & Finkbeiner (2011) extinction maps.However, the source is indeed located in a dense and dark part of the nebula, and a total extinction A V ∼ 100 mag is not unrealistic.Such high extinction is also confirmed by the spectral energy distribution (SED) of the point source, shown in Fig. 3, which is rather extreme (see Appendix C for details about the SED data).We tried to fit the SED using a variety of stellar spectra libraries available at VOSA (Bayo et al. 2008), andRobitaille et al. (2007) and Baraffe et al. (2015) stellar models, however the results are inconclusive since none of the models resulted in a reasonably good fit to the data.For instance, adopting the aforementioned extinction and distance, the SED models suggest this source is likely to be a massive YSO.Limiting ourselves to the range 2 µm to 22 µm (K s from VVV plus W1 to W4 from WISE) the SED can be fitted to the Robitaille et al. (2007) model grid with a deeply embedded, class 0 or class I YSO model, but model parameters should be treated with caution due to the presence of a massive cloud core that dominates the far infrared SED and the unusual nature of the system.The difficulty in matching the SED with YSO models arises because in the archival data the photometry was taken at different epochs, thus at different phases, and with different apertures (see Appendix C).
The VVV images of this region show that the vicinity resembles an active star formation region (SFR), being rich with interstellar clouds, compact cores, YSOs, and young star clusters, as example of DB121 and DB123 which have distances measured by Soares et al. (2008) as D = 1.6 kpc.Assuming the same distance for VVV-WIT-12 and its nebula yields a size of about 0.2 pc for the whole illuminated nebular region (30 ′′ ).While this is a safe assumption, the confirmation of the distance needs further observations.
Archive search at the VVV-WIT-12 position revealed additional measurements at longer wavelengths.In particular, (NEO)WISE data (Wright et al. 2010;Mainzer 2).Such long period is toward the upper end of what is seen in LPVs, and we note that YSOs can also be periodic too.The amplitude is variable in the K s -band based on 3 cycles.The first and second cycles show observed peak-to-peak amplitudes of A Ks ∼ 1.4 and 1.7 mag, respectively.
The surrounding nebula changes its brightness and J − K s color during the period of observations (Fig. 4, see details in Appendix A).In year 2015 the Northern part appears redder than in 2010, while the opposite occurred with the Southern part, which became bluer.Scattered light from two regions located on bright nebular patches at opposite sides of VVV-WIT-12 shows the same periodic behavior as the central source, confirming that the scattered light is indeed an echo of the source (see Fig. 5).The red curve shows a substantial lag (about half a period) which may indicate that it is caused by back scattering off dust more distant than the source.In this case, if we take the 2.1 year lag as the light travel time from the source to the back side of the nebular cavity, it would yield a distance of about 0.32 pc.On the other hand, the blue curve is in phase with central source, which is interesting since the forward scattered light also has to travel an extra distance compared to the direct light.However, this consideration only holds for single scattering.Because of the dense environment and the large extinction toward the source, this will likely not be the case here.Perhaps the blue region has a lower optical depth so that the light can more easily leak toward us.
The scale of the variations across the nebula is relevant, and we take 30 ′′ as a representative scale, fully contained within Fig. 1, that covers 1.1 ′ on a side.We use 30 ′′ because this is the total size of the region that varies in brightness and color.Assuming a distance D = 1.6 kpc, 30 ′′ corresponds to 0.24 pc.In this case, the fluctuating region would be 40 times smaller than the size of the Orion nebula for example.The lighttravel time throughout a region of this size would be about 0.8 yr.If the source is more distant, e.g., located within the bulge at D = 8 kpc, then this region would be larger, 30 ′′ = 1.2 pc.This is a large scale for brightness variations, considering that it cannot be attributed to gas motions because that would require superluminal velocity.Instead, the preferred explanation is that the nebula changes as it is being illuminated by the variable central source.We estimated the range of luminosity variation for the point source using the lowest and highest values of fluxes in the NIR/MIR bands, assuming the distance of 1.6 kpc and no extinction correction.This yields ∆L = 657 ± 71 L ⊙ and 800 ± 56 L ⊙ , where the errors were derived by using the photometric errors as lower/upper bounds.
Spectroscopic follow-up of the central source (VVV-WIT-12) was secured on June 11 2023 (JD 2460106, corresponding to ϕ = 0.17 cycles according to the ephemeris used to phase fold our light-curves) with the  2).This shows that the source is very red, with a featureless steadily rising continuum, lacking strong near-IR emission lines.
Data appearing in the SED are described in Appendix C.
TripleSpec spectrograph1 at the SOAR Telescope (Wilson et al. 2004;Herter et al. 2008).The spectrum (Fig. 3) shows that the source is very red, with a featureless steadily rising continuum, lacking strong near-IR emission lines.The feature seen at ∼ 2.0 µm is the residual of the removal of a very strong telluric line.Such featureless but very red spectra are typical of some YSOs (Contreras Peña et al. 2020;Guo et al. 2022).However, a very dusty Mira cannot be discarded, as they may also exhibit a similar spectral behaviour.

DISCUSSION
As with most VVV-WIT objects discovered, we still do not know what is this.Not only the source is variable, but also the surrounding nebulosity changes brightness and shape.Given the long period and the classification as an LPV, initially we considered two possibilities: (i) an LPV that illuminates the encircling ISM inhomoge-Figure 5. Variability of the scattered light from two regions of the VVV/VVVX Ks images.Left: the emission was integrated inside the circular blue and red masks.A smaller mask was also placed on the object.The image orientation is the the same as in Figs. 1 and 4. Right: the resulting light curves for the three regions covering the same epochs as Fig. 2. All curves show the same periodicity behavior, but the red curve presents a substantial lag with respect to the point source, while the blue curve is in phase with VVV-WIT-12.Time in the x-axis is shown in years (top) and MJD (bottom).
neously through thick dust clouds in its atmosphere, or (ii) an LPV that is coming out from behind an unrelated interstellar cloud.
The LPV hypothesis is discarded by examining the magnitudes.The expected magnitude for an LPV with P = 1525 days is M Ks ∼ −11 mag (e.g., Macri 2017).If the object is located at D = 1.6 kpc, its distance modulus would be m − M 0 = 11.02mag.Therefore its apparent unreddened magnitude should be about 0 mag, which is inconsistent with the observed magnitude of K s =13.2 mag.A possible exception might have a rare case of J type Carbon star, that are fainter than other AGB stars, reaching M K ∼ −2.5 mag (Abia et al. 2020).However, these objects are rather blue, inconsistent with the observed colors and the spectral energy distribution that is rising to the near-IR.
We also note that the observed geometry for the nebular changes is inconsistent with the spherically symmetric picture that a light echo would produce, as seen for example in historical SNe remnants and η Carina (Rest et al. 2005).This suggests that the dust configuration surrounding VVV-WIT-12 causes a light echo which is by far more complex than those of SN.In principle, the dust distribution could be characterized based on the phase lag of the scattered light, but this detailed 3D modeling is beyond the scope of this paper.
Considering now that VVV-WIT-12 is a YSO because of its luminosity and spectrum as well as the long term oscillatory pattern, as seen in other YSOs in the midinfrared (e.g., Park et al. 2021, and references therein), attributed to either variable extinction or a cyclical change in accretion rate, there are a few other alternatives to explain the observations: (i) the light echo from a periodic YSO, or (ii) the progressive illumination of the ISM by the precessing disk of a YSO, or (iii) an orbiting clump of material that blocks the light path of the variable source.Even though at present the YSO option remains as the most reasonable hypothesis, we are unable to confirm this without further observations.

CONCLUSIONS
We have discovered VVV-WIT-12, a very reddened (A V > 100 mag) periodic variable source with P = 1525 days, that has a smooth red continuum and that is located inside a nebula that changes in brightness and color with time.The real nature of the source is still unknown, but based on the extensive near-IR imaging and spectroscopic observations, our preferred explanation is a periodically variable YSO with a light echo.
In particular, massive light echo searches are being planned for the Vera C. Rubin Observatory (Li et al. 2022), taking advantage of a median of 815 images scheduled for every position of the sky, reaching g ∼ 24.5 mag, r ∼ 24.0 mag.These searches will use artificial intelligence tools like automated deep convolutional neural networks to make many important light echo discoveries from past SNe, especially across the Milky Way plane (Rest et al. 2005).Our work suggests that these massive light echo explorations may be contaminated by variable objects like the one presented here.However, these light echo searches will allow to discover and classify VVV-WIT-12 analogs, and to study their time evolution.Also, not less important, these searches would open the door for other serendipitous discoveries.
In conclusion, VVV-WIT-12 appears to be a very rare long period YSO that acts as a cosmic beacon flashing its surrounding nebula.This is worthy of further study, and suggested future observations are near-IR spectroscopy of the nebula to confirm its nature, and more continuous monitoring of the variability of this interesting object.
We gratefully acknowledge the use of data from the ESO Public Survey program IDs 179.B-2002 and 198.B-2004  To inspect the variability of the scattered light across the surrounding nebula, we analyze the VVV/VVVX images taken at different epochs.In a first step, we selected K s -band images taken near to maximum/minimum of the VVV-WIT-12 light-curve, according to the P = 1525 ephemeris: a K s -band image taken on April 18, 2010, at phase ϕ ∼ 0.75; and a K sband image from August 15, 2015, at phase ϕ ∼ 0.05.Then we divided the 2010 image by the 2015, creating a composite K s -band image with the differences.As presented in Fig. 4, the nebula changes in brightness, appearing brighter in the first epoch, when the VVV-WIT-12 is closer to the maximum (ϕ ∼ 0.75).We then repeated the same steps as above using J and K s -band images taken in sequence at those nights (in 2010 and 2015).A composite J/K s image was created by dividing the J-band by the K s -band, and finally dividing the 2010 J/K s image by the 2015 J/K s .This shows the spatial variation of the near-IR nebular color, which cover different places than the brightness variations.After all the divisions, the net differences both in brightness and color are ∼ 1 − 2%, but consistent across all the nebula.In Fig. 4 the color scale has been stretched to emphasize the changes.
Since our first analysis using two epoch showed variations in brightness and color across the nebula, we made use of a set of K s -band images taken along the VVV/VVVX campaign, ranging from from April 2010 to August 2019.We selected two regions located on bright nebular patches at opposite sides of the point source to place two circular masks of 6 pixel radius in order to integrate the emission, after subtracting the background.A third and smaller mask of 4 pixel radius was also placed on the position of the point source (VVV-WIT-12), for simultaneous monitoring (see the left panel of Fig. 5).The extracted magnitudes (instrumental, i.e., not photometrically calibrated) were combined with the MJD to create a resulting light curve for both masks placed on the nebula simultaneously with the central point source.In extracting the magnitudes, the uncertainty is in the range of σ Ks ≈ 0.2 mag.The light curves for the two regions (blue and red) were arbitrarily shifted to approximately match the mean level of the point source (black light curve), as presented in the right panel Fig. 5.All curves show the same periodicity behavior, but the red curve presents a substantial lag with respect to the point source, while the blue curve is in phase with VVV-WIT-12.

B. VIRAC2 K S -BAND LIGHT-CURVE
Here we present all the VIRAC2 photometric measurements of the central source VVV-WIT-12 used to build the light curve presented in Fig. 2.These data listed in Table 1 comprise 203 K s -band data-points.The source is relatively bright, and the typical photometric errors are very small (∼ 0.02 mag).

Figure 1 .
Figure 1.Composite VVV-WIT-12 Ks-band image made using three different epochs: years 2010 (red), 2011 (green), and 2012 (blue).This is the discovery image for VVV-WIT-12, that shows the variation of the central source as well as the nebula.The image is 67 ′′ on a side, centered in VVV-WIT-12 at RA/DEC (J2000) = 17:17:20.29,−36:08:43.9(l, b = 350.7080,+1.0259 deg), and oriented along Galactic coordinates.Sources that are non-variable in the near-IR should look white, while variable sources are colored differently.The point source was brighter in 2010 and the nebula changed brightness in different parts along the epochs.

Figure 2 .
Figure 2. Left panel: Light curve of VVV-WIT-12 covering years 2010 to 2022.Grey points represent the Ks-band observations from VVV/VVVX, and cyan points represent the W1-band observations from WISE (arbitrarily shifted by 3 mag).Right panel: VVV-WIT-12 light curves phased using P = 1525 days.The different cycles are color-coded as labelled to show the amplitude changes.While the minimum light seems to be the same in the two cycles observed with the Ks-band, the maximum is about 0.3 mag brighter in the second cycle, indicating a possible eruption.et al. 2011) covers a longer baseline than VVV, from March 12 2010 to August 21 2022, however with a sparser cadence.The W1-band light curve presents a periodicity of ∼ 4 years with an amplitude A W 1 = 1.1 mag, in agreement with the VVV observations (see Fig.2).Such long period is toward the upper end of what is seen in LPVs, and we note that YSOs can also be periodic too.The amplitude is variable in the K s -band based on 3 cycles.The first and second cycles show observed peak-to-peak amplitudes of A Ks ∼ 1.4 and 1.7 mag, respectively.The surrounding nebula changes its brightness and J − K s color during the period of observations (Fig.4, see details in Appendix A).In year 2015 the Northern part appears redder than in 2010, while the opposite occurred with the Southern part, which became bluer.Scattered light from two regions located on bright nebular patches at opposite sides of VVV-WIT-12 shows the same periodic behavior as the central source, confirming that the scattered light is indeed an echo of the source (see Fig.5).The red curve shows a substantial lag (about half a period) which may indicate that it is caused by back scattering off dust more distant than the source.In this case, if we take the 2.1 year lag as the light travel time from the source to the back side of the nebular cavity, it would yield a distance of about 0.32 pc.On the other hand, the blue curve is in phase with central source, which is interesting since the forward scattered light also has to travel an extra distance compared to the direct light.However, this consideration only holds for single scattering.Because of the dense environment and the large extinction toward the

Figure 3 .
Figure 3. Top: spectral energy distribution (SED) of VVV-WIT-12 covering 1 − 1000 µm, obtained with the Vizier photometry viewer a .Solid circles mark the data points from VVV and WISE within the range 2 µm to 22 µm (see Section 2).Bottom: near-IR spectrum of the VVV-WIT-12 stellar source, acquired with the TripleSpec spectrograph at the SOAR telescope in Chile, in 2023 June 11 (JD 2460106, corresponding to ϕ = 0.17 cycles, see the right panel of Fig.2).This shows that the source is very red, with a featureless steadily rising continuum, lacking strong near-IR emission lines.

Figure 4 .
Figure 4. Top: Composite Ks-band image made using a year 2010 image (with VVV-WIT-12 near phase ϕ ∼ 0.75 cycles) divided by a year 2015 (ϕ ∼ 0.05 cycles).The nebula varies in Ks-band, appearing brighter in the first epoch.Bottom: Composite J/Ks image showing the spatial variation of the near-IR nebular color.Also, the color variations cover different places than the brightness variations.In both panels the field of view is 67 ′′ on a side, oriented along Galactic coordinates, with longitude increasing to the left and latitude towards the top, and covers the same field as Fig. 1.The color scale has been stretched to emphasize the changes.A horizontal bar shows the relative intensity in each case.
taken with the VISTA telescope and data products from the Cambridge Astronomical Survey Unit.This publication makes use of data products from the Wide-field Infrared Survey Explorer, which is a joint project of the University of California, Los Angeles, and the Jet Propulsion Laboratory/California Institute of Technology, funded by the National Aeronautics and Space Administration.D.M. gratefully acknowledges support from the Center for Astrophysics and Associated Technologies CATA by the ANID BASAL projects ACE210002 and FB210003, by Fondecyt Project No. 1220724, and by CNPq/Brazil through project 350104/2022-0.R.K.S. acknowledges support from CNPq/Brazil through projects 308298/2022-5 and 350104/2022-0.ZG is supported by the ANID Fondecyt Postdoctoral program No. 3220029.ZG acknowledges support by ANID, -Millennium Science Initiative Program -NCN19 171.The work of FN is supported by NOIRLab, which is managed by the Association of Universities for Research in Astronomy (AURA) under a cooperative agreement with the National Science Foundation.CM acknowledges support from the UK's Science and Technology Facilities Council (ST/S505419/1).