Over the course of these past three years, efforts have been made to better classify the Mid-IR sources selected by Acker et al. (2016) (e.g. optical spectra). These objects were selected based on their Planetary Nebula (PN)-like colours in IRAS and WISE data, and are meant to represent a preliminary method to detect compact PNe. However, as mentioned in their work, further photometric and spectroscopic studies are needed to better assess the nature of these objects. This is because many “mimics” display similar colours in IRAS and WISE, most notably AGNs, HII regions, Post-AGB stars and YSOs (Acker et al., 2016). The article was also published in the popular French astronomy magazine Astronomie. (February 2016 edition). Based on photometric data and imagery from various other [modern] surveys (e.g. DECaPS, SHS and Herschel), this article shows that the PN-IR objects Mul-IR 56, Mul-IR 58 and Mul-IR 62 may have good chances of being true Planetary Nebulae. I also feature several other PN-IR objects at the end of this work that I am currently in the process of studying further.
Figure 1: Cover image of the article published in the February 2016 edition of the Astronomie magazine (written by Lionel Mulato). Image credit: Astronomie/SAF
1. IRAS and WISE Properties of PNe
At wavelengths inferior to 50 microns, compact PNe, YSOs (Class 0 and I), HII regions and Proto-PNe may appear quite similar (see Acker et al., 2016). However, at higher wavelengths, these often tend to differ (especially YSOs and HII regions). Indeed, these often display Far-IR excess (caused by their dusty envelope/environment) while PNe generally do not. Indeed, PNe tend to have peak emissions in the Mid-IR, around 20-50 microns (e.g. figure 2a) while YSOs and HII regions peak in the Far-IR (e.g. figure 2b). Spectral Energy Distribution (SED) plots can thus be good ways of classifying objects, as can be seen on the Zooniverse/Disk Detective website. As a result, one can simplify the morphology these SEDs into a set photometric criteria (e.g. flux ratio) in order to classify such objects. Such work has notably been done using the various IRAS IR filters (see figure 2).
Figure 2: Plot comparing IRAS photometric properties (flux density ratios) of numerous confirmed PNe, and PN-IR objects from (Acker et al., 2016). Despite the dispersion, note how most PNe are defined within a single “photometric zone”. However, a very large amount do not meet these criteria. Image credit: Acker et al. (2016)/Astronomie Feb 2016.
Indeed, Pottasch et al. (1988) and Acker et al. (2016) based their selection on the following predefined Flux Density ratio values: 12/25 > 0.5 and 25/60 < 0.35. Thus, these criteria describe objects with a strong slope between 12 and 25 microns, and relatively weak or descreasing slopes between 25 and 60 microns, hence reflecting the photometric properties of most PNe as shown by their SEDs (e.g. figure 3). The small IRAS 12/25 Flux ratio displayed by most PNe can be translated by their low WISE W3/W4 (12/22 micron) ratio. However, based on the SED morphology of PNe (e.g. figure 3), one could also use the WISE W1/W4 ratio as a substitute, as this is roughly/graphically proportional to the WISE W3/W4. Consequently, in AllWISE W1+W2+W4 images (Colour-coded Blue-Green-Red), most PNe appear as large “red spots” reflecting their Mid-IR excess, in comparison to their relatively weak Near-IR signal (see figure 4).
Figure 3: SED of MR 22 (a bright PN) as derived from on various astronomical catalogues calculated by CDS/SIMBAD. Notice how the continuum peaks around 70 microns, with a steep slope from the Near-IR to the Mid-IR. Image credit: CDS/SIMBAD.
Figure 4: AllWISE Image extracts comparing various PN-IR objects. All the objects presented in this image are of the List-I, with the exception of Mul-LDû-IR 16. Mul-IR 9 and Mul-IR 14 are confirmed PNe. Image credit: Aladin Lite/AllWISE.
Note: The coordinates for all the PN-IR objects can be found here, as well as in the article by Lionel Mulato.
2. The Herschel-WISE Diagram
In regards to the work by Acker et al. (2016), all PN-IR objects (List I and II) were initially selected based on their relatively strong response in the W4 filter, with those of the List I all ressembling the objects shown in figure 4 (with the exception of Mul-LDû-IR 16, a list II object). Mul-IR 56, Mul-IR 58 and Mul-IR 62 are all members of the List I. Some object of the List II however have a significant response in the W1 and W2 bands, relative to W3 and W4, such as Mul-LDû-IR 16. The team then sorted/retained the WISE-selected objects according to their IRAS properties. However, in comparison to many modern IR surveys (e.g. Herschel and WISE), IRAS data is of relatively low resolution. Consequently, I initially decided to recreate the IRAS diagram (Figure 7) according to photometric bands from more modern surveys. Firstly, I exchanged the IRAS 12/25 ratio by the WISE W1/W4 (see explanation above), and I chose the PACS-Blue/W4 ratio as a [mathematically inverse] substitution for the IRAS 25/60. The PACS-Blue filter corresponds to the 70 micron band used by the Herschel telescope. These values were provided by the Herschel Point Source Catalogue.
Figure 5: A projection of the Herschel FOV. Notice how this is focused at the very center of the galactic plane and other areas rich in star formation. Image credit: ESA/Herschel.
In this search I focused on finding known objects in AllWISE images that dispayed the same colours as those in figure 4, located within the Herschel FOV. I also selected YSOs with similar properties from the HOPS catalogue and Class 0 YSOs from Savadoy et al. (2014). As a result, my preliminary results contain 44 PNe (including PN candidates), 41 YSOs, 12 HII compact regions and 5 Proto-PNe (PPNe). Since the Herchel FOV mostly focused on obscured starforming regions, AGNs were excluded as these were relatively unlikely to be present. PPNe were difficult to find in the Herschel FOV, hence the small number of subjects presented here. In the case of HII regions however, it is possible that future work could include numerous more of these. Furthermore, note that corrections due to magnitude/flux extinction have not been made. While interstellar reddening is unlikely to significantly affect the PACS70 and the W4 band flux values, the W1 band fluxes may be significantly underestimated in the case of highly obscured sources. Moreover, it is possible that the Herschel flux values do not correspond to flux density, as I initially thought. WISE Flux density values were calculated according to the following protocol. In order to further verify the accuracy of the WISE calculations, I made comparisons with other works (e.g. figure 6 and 9).
Figure 6: Test results in the case of well known edge-on SEDs, MY Lup. Note how both studies yield similar results. Note the difference in data in at optical wavelengths is most likely due to the variable nature of these objects. Data: 2MASS, Spitzer, GSC2.3, IRAS, AKARI, UCAC4, WISE.
Despite the statistically small number of samples, it seems possible to distinguish certain “areas” that correspond more to PNe rather than YSOs and HII regions (despite many PNe displaying properties that overlap with the latter two classes). Too few PPNe are available to draw any definite conclusion in regards to any “photometric zones”, especially as the dispersion is large, and three or four of the samples display properties that fall within the “PN region”. Interestingly, the HII regions selected here display relatively similar properties (poor dispersion), but more subjects are required to define a definite “zone”, if possible. As expected, YSOs of PN-like WISE colours have an average PACS70/W4 slope that is higher than for PNe (at least the ones chosen here). Visually, I find that PNe tend to roughly have PACS70/W4 values inferior to 0.5. The HII regions calculated here interestingly all appear to have a PACS/W4 < 0.5, which is coherent with the IRAS 25/60 ratio for these objects (see figure 2), considering the PACS-Blue values are measured in flux density. In addition, despite not being present in figure 7, the object Mul-IR 9 (now a confirmed PN) is also located in the “PN region”.Figure 7: Plots displaying the Herschel-WISE photometric properties of Mul-IR 56 and Mul-IR 58 compared to 44 Planetary nebula (PN) candidates, 41 confirmed Young Stellar Objects (YSOs), 12 HII regions and 5 PPNe of the Herschel Orion Protostar (HOPS) catalogue, and likely Class 0 YSOs from Sadavoy et al. (2014). Most HII regions and YSOs were selected based on their WISE colours being similar to those commonly observed PNe (and hence the IR objects of the Acker et al., 2016 database).
In the case of Mul-IR 56 and Mul-IR 58, these display photometric properties that best match those of PNe, especially Mul-IR 58. However, Mul-IR 56 is close to the “area” where both PNe and YSOs overlap. The few HII regions studied here have a significantly higher PACS70/W4 ratio than Mul-IR 56 and Mul-IR 58. Mul-IR 62 is not part of the Herschel Point Source Catalogue, but does however appear in the Herschel FOV (see below).
However, there might be a trend for which the PACS70/W4 ratio of YSOs decreases with decreasing W1/W4 ratio. This would be coherent with the decrease in the IR slopes as the YSOs progressively accrete/loose their outer envelopes (see figure 7). Interestingly, this property appears to cause YSOs to photometrically distinguish themselves from PNe, despite more evolved YSOs having similar PACS70/W4 ratios to the PNe chosen here. In support of this hypothesis, most of the Class 0 YSO candidates (from Savadoy et al, 2014) selected here have the highest PACS70/W4 ratios and overall the smallest W1/W4 ratios. On the other hand, more evolved YSOs, such as Mul-LDû-IR 16 (see future work), are located in the zone with the lowest IR slopes.
Figure 8: Composite image explaining the link between the photometric properties of YSOs, and their evolution. Indeed, as the YSO and its circumstellar matter accretes material, the outer envelope progressively vanishes, leading to a progressive decrease in the mid-IR excess. This leads to a decrease in the IR slopes. Image credit: Lada et al. (2002).
However, it is important to note that the Infrared flux values in the Near-IR/Mid-IR region may be significantly influenced by the inclination of the YSO according to our perspective (Robitaille et al., 2006). More specifically, as seen edge-on, YSOs may be relatively obscured between 1 – 15 microns (e.g. HOPS 136 [Fischer et al., 2014], MY Lupi [see figure 6]), in comparison to a YSO observed nearly pole-on (e.g. HOPS 329 [Manoj et al., 2013]). To demonstrate this effect, see figure 9. HOPS 329 flux density values were compared/verified to those of Manoj et al. (2013).
Figure 9: Comparison between the SED of a nearly pole-on YSO (HOPS 329, incorrectly labelled HOPS 392) and an edge-on YSO (HOPS 136). Note how the latter SED shows a strong and broad anomaly (dip) around 10 μm, while the former shows a smooth trend in this region. Image credit: Université Grenoble Alpes – Trygve Prestgard
3. Follow-Up Infrared Observations
For a more comprehensive view of the studied PN-IR candidates, I chose to calculate/extract the flux density values from various other Infrared catalogues listed in Vizier. As can be seen in figure 10, Mul-IR 56 appears to have a SED that displays peak Infrared emissions between 30 and 60 microns, very much like a PN. On the other hand, despite the PN-like WISE-Herschel colours for Mul-IR 58 (see figure 7), AKARI and IRAS give values that differ significantly from Herschel, making it difficult to say whether the peak IR emissions lay at wavelengths smaller or larger than 100 microns. In the case of Mul-IR 62, no values were provided by the Herschel Point Source Catalogue. The only flux density values beyond 50 microns are provided by AKARI and IRAS (60 and 90 microns). With only two values (and IRAS being of relatively low resolution), it is difficult to say whether or not the peak IR emissions of Mul-IR 62 lie below 100 microns or not. However, if the values are considered accurate, the plot may be suggestive of a peak near 50 microns.
Figure 10: SEDs plotted for Mul-IR 56 (top), Mul-IR 58 (middle) and Mul-IR 62 (bottom), based on flux values extracted/calculated from various public survey data available via Vizier. While Mul-IR 56 has a SED typically expected of a PN, Mul-IR 58 and Mul-IR 62 are a bit more difficult to interpret based on these data alone, but may be suggestive of PNe too.
Fortunately, The SIMBAD/CDS portal allows access to numerous different combinations of images, including a coloured combination of the three Herschel/PACS filters (70 + 100 +160 microns). With the images being colour-coded Blue (70 microns), Green (100 microns) and Red (160 microns), PNe should appear distinctively blue due to the morphology of their SED. These images can hence serve as a visual substitue for the lack of values beyond 50 microns. Interestingly, all three objects display a distinctively blue colour in these images, with Mul-IR 62 even appearing somewhat elongated. This is hence further evidence in regards to their PN nature.
Figure 11: Herschel colour images showing Mul-IR 56, Mul-IR 58 and Mul-IR 62 compared to PNe and some YSOs. Notice how the former display a distinctive blue colour (reflecting the decrease in the emissions beyond 70 microns), contrary to the YSOs. Image credit: SIMBAD/Herschel_RGB.
4. Follow-Up Optical Observations
Since early 2018, high-resolution images from the DECaPS survey were released to the public. These images allowed the discovery of numerous new PNe that were otherwise too faint, or too stellar, to “stick out” in optical images (LDû et al., 2019). Various examples of PNe and candidates (as seen in DECaPS) are shown in figure 12. Notice how the apparent colours of these nebulae vary depending on the intensity interstellar reddening.
Figure 12: DECaPS image extracts of six PNe and candidates suffering various degrees of interstellar reddening. Notice how Mul-IR 56 and Mul-IR 58 display the colours of highly reddened PNe. Image credit: Aladin Lite/DECaPS.
As can be seen in figure 13, all three Mul-IR objects display traces of nebulosity and colours typical of highly reddned PNe. Moreover, notice how Mul-IR 58 and Mul-IR 62 appear to display a possible bipolar morphology. The nebulous morphology shows that these are hence unlikely to be YSOs.
Figure 13: DECaPS Image extracts showing Mul-IR 56, Mul-IR 58, and Mul-IR 62 too scale. Notice how Mul-IR 56 and Mul-IR 62 appear to possibly be bipolar, while Mul-IR 58 appears much more compact and elliptical in morphology. Note also that the colours match those of the highly reddened PN candidates displayed in figure 2. Image credit: Aladin Lite/DECaPS.
In addition to their nebulous signature in DECaPS, traces of the nebulosity can be found in SHS (SuperCosmos Halpha Sky Survey) images, in the case of Mul-IR 56 and Mul-IR 62. More importantly however, these two nebulae appear brighter in the Halpha plates in comparison to those taken in the r filter (see figure 14). This suggests that the nebulae display Halpha emissions, very typical of PNe. One could expect that the apparent emissions in Halpha may be stronger had the nebulae not been obscured by Interstellar matter. Mul-IR 58 is likely too obscured and/or too faint to display any significant Halpha emissions, if indeed a true PN. However, let us note that PPNe do not emit significantly in Halpha (these are reflection nebulae, unlike PNe), hence one could technically consider the latter to be a possible Post-AGB/PPN object. Note however that HII regions and YSOs may also display Halpha emissions, but the IR colours explained earlier do not appear in favour of this classification.
Figure 14: SHS Short Red (left) and SHS Hα (right) image extracts showing Mul-IR 56 and Mul-IR 62. Notice the nebulous nature of these objecst in Hα, while they appears to be poorly resovled and alot fainter in SHS Short Red. Image credit: SuperCosmos Halpha Survey.
Based on this small study of Optical and IR photometric data/imagery, it is possible to conclude that Mul-IR 56, Mul-IR 58 and Mul-IR 62 may in fact be true PNe. Firstly, the Herschel-WISE colours of Mul-IR 56 and Mul-IR 58 fall within the “photometric region” of PNe, rather than YSOs or HII regions. Furthermore, the SEDs of Mul-IR 56 and Mul-IR 62 are rather suggestive of a PN (especially the former). The SED of Mul-IR 58 is difficult to interpret beyond 60 microns due to the high dispersion of flux values. However, the Herschel Image colours seem suggestive of PN-like continuum emissions beyond 70 microns.
DECaPS images show all three objects to be nebulous in nature, which is strong evidence against these being YSOs. The nebulous nature of Mul-IR 56 and 62 is also apparent in SHS Halpha images, with the latter also being nebulous in Herschel too. The SHS images show that Mul-IR 56 and Mul-IR 62 are very likely to be emission nebulae, as expected of PNe. However, if Mul-IR 58 is indeed a true PN, the object is likely too obscured for the Halpha signals to be detected. Hence, the lack of apparent Halpha emissions may also technically include the nebula as a PPN candidate (as WISE-Herschel PPN colours seem to overlap with those of PNe).
However, one must note that the study is preliminary and the number of reference samples are statistically low. Only detailed spectra can reveal the true nature of these three objects. Unfortunately, with these nebulae being faint and heavily obscured (and Mul-IR 62 partially eclipsed by a foreground star), optical spectra may be difficult to obtain for these objects, hence the importance of finding reliable IR diagnostic criteria.
Future and Ongoing Work
Studying the variability of sources using DSS and Pan-STARRS1 image plate comparison, as well as ASAS-SN can “weed out” some YSOs and Late-type stars from our list of PN candidates. Results: Mul-LDû-IR 16, Mul-IR 20, Mul-IR 64 and Mul-IR 92 are all clearly variable. Mul-LDû-IR 16 is most likely a YSO. Mul-IR 21 may be some evolved star (Gomez et al., 2015), while Mul-IR 64 displays Long Period Variations (LPV) typical of PPNe. In fact, Mul-IR 64 has already been identified as variable by the Bochum survey, and is listed as an unclassified variable according to the AAVSO/VSX (GDS_J1633219-461042). Mul-IR 92 is currently listed as a possible YSO by the AAVSO/VSX.
Figure 15: ASAS-SN light curve showing the variable nature of Mul-IR 64. The variations are typical to those of a Semi-regular star, with a period of ~25 days. Based on the object’s optical and IR colours (“yellow” star with Mid-IR excess), in addition to the light curve, it is most likely a PPN.Image credit: ASAS-SN.
Mul-LDû-IR 3: A likely an elliptical Planetary Nebula, and has independently been listed as a PN by Gomez et al. (2015). The object is bright and clearly elliptical in DECaPS images (see figure 16), with apparently strong Halpha emissions.
Figure 16: DECaPS image extracts of Mul-IR 56, Mul-IR 62 and Mul-LDû-IR 3 showing their nebulous nature. In comparison to the Mul-IR objects, Mul-LDû-IR 3 is much less reddened and elliptical in morphology.Image exctract: Aladin Lite/DECaPS.
A few other sources (e.g. Mul-IR 31 and Mul-IR 65) may perhaps be stellar PNe based on their optical colours (Optically blue with likely Halpha emissions, as seen in Pan-STARRS, DECaPS and SHS) and their IR properties. If I remember correctly, ASAS-SN data did not reveal any variability in the case of Mul-IR 31, hence another factor that increases the chances of this object being a true PN! 🙂
Figure 17: Pan-STARRS1 and DECaPS image extract of Mul-IR 31 and Mul-IR 65. Both objects are stellar in appearance, but their otical and IR colours may be in favour of them being true PNe. Work in progress! Image credit: Aladin Lite/DECaPS and Pan-STARRS1.
Future work should also include a better study of Spitzer data, as PNe tend to display particular colours in these images, that most often distinguish them from HII regions and many YSOs. In the Aladin/lite Spitzer/IRAC images these tend to appear “green” (best emissions in the IRAC2 filter(?)). Note that Mul-IR 56, Mul-IR 58 and Mul-IR 62 all display this green colorimetry. Mul-LDû-IR 3 is located outside the Spitzer/IRAC FOV.
Figure 18: Spitzer Image extracts showing Mul-IR 56, Mul-IR 58 and Mul-IR 62. Notice how Mul-IR 56 and Mul-IR 62 are both clearly nebulous in these images, however the reoslution is likely to small to display their bipolar nature. The reolution is likely also too low to display the nebulous nature of Mul-IR 58. Image credit: Aladin Lite/Spitzer.
CDS: Simbad, Vizier (Université de Strasbourg).
Fischer, W. J. et al. (2014) ‘HOPS 136: AN EDGE-ON ORION PROTOSTAR NEAR THE END OF ENVELOPE INFALL’, The Astrophysical Journal, 781(123), pp. 11.
Gómez, J. F.; Rizzo, J. R.; Suárez, O.; Palau, A.; Miranda, L. F.; Guerrero, M. A.; Ramos-Larios, G.; Torrelles, J. M (2015): ‘A search for water maser emission toward obscured post-AGB star and planetary nebula candidates’ Astronomy & Astrophysics, Volume 578, id.A119, 15 pp.
Le Dû, P. (2019) ‘Nebuleuses planétaires, Découvertes & Confirmations ‘, Astronomie
Manoj, P. et al. (2013) ‘HERSCHEL/PACS SPECTROSCOPIC SURVEY OF PROTOSTARS IN ORION: THE ORIGIN OF FAR-INFRARED CO EMISSION’, The Astrophysical Journal, 763(2)
Mulato, L. (2016) ‘Nébuleuses dans L’Infrarouge Moyen’, Astronomie, Feb 2016
Pottasch, S. R., Bignell, C., Olling, R., Zijlstra, A. A (1988) ‘Planetary nebulae near the galactic center’
Robitaille, T., Whitney, B., Wood., K. (2006) ‘Interpreting Spectral Energy Distributions from Young Stellar Objects. I. A Grid of 200,000 YSO Model SEDs’
Sadavoy S. I. et al. (2014) ‘CLASS 0 PROTOSTARS IN THE PERSEUS MOLECULAR CLOUD: A CORRELATION BETWEENTHE YOUNGEST PROTOSTARS AND THE DENSE GAS DISTRIBUTION’, The Astrophysical Journal, 787(L18), pp. 6.
Spitzer Science Center (2019) Magnitude/Flux Density Converter: Point Sources, Available at: http://ssc.spitzer.caltech.edu/warmmission/propkit/pet/magtojy/