No evidence has been found that it is part of a multiple star system. Its great infrared (IR) excess makes it one of the brightest objects in the local part of the galaxy at wavelengths of 5 to 20 microns (µm) and indicates a dust shell or heated disk.[14][15] It is about 17±8 times the mass of the Sun (M☉). It is surrounded by a complex asymmetric circumstellar envelope (CSE) caused by its mass loss. It produces strong molecular maser emission and was one of the first radio masers discovered. VY CMa is embedded in the large molecular cloudSh2-310, a large, quite local star-formingH II region—its diameter: 480 arcminutes (′) or 681 ly (209 pc).[16][17]
The radius of VY CMa is very roughly 1,420 times that of the Sun (R☉), which is close to the modelled maximum, the Hayashi limit, a volume nearly 3 billion times that of the Sun.[3] Taking this mid-point estimate as correct, an object travelling at the speed of light would take 6 hours to go around its surface, compared to 14.5 seconds for the Sun.[18] If this star replaced the Sun its surface would, per this approximation, be beyond the orbit of Jupiter.[3]
The first known-recorded observation of VY Canis Majoris is in the star catalogue of the French astronomer Jérôme Lalande in 1801,[b] which lists it as a 7th order of magnitude star. Further quite frequent studies of its apparent magnitude imply the light of the star as viewed from Earth has faded since 1850, which could be due to emission changes or a denser part of its surrounds becoming interposed (extinction).[19]
Since 1847, VY Canis Majoris has been described as a crimson star.[19] During the 19th century, observers measured at least six discrete components, suggesting that it might be a multiple star. These are now known to be bright zones in the host nebula. Observations in 1957 and high-resolution imaging in 1998 all but rule out any companion stars.[19][20]
The variation in the star's brightness was first described in 1931, when it was listed (in German) as a long-period variable with a photographic magnitude range of 9.5 to 11.5.[28] It was given the variable star designationVY Canis Majoris in 1939, the 43rd variable star of the constellation Canis Major.[29]
Map of the giant molecular cloud Sharpless 310 and its surroundings
Combining data from the mentioned telescope with others from the Keck in Hawaii it was possible to make a three-dimensional reconstruction of the envelope of the star. This reconstruction showed that the star's mass loss is much more complex than expected for any red supergiant or hypergiant. It became clear that the bows and nodules appeared at different times; the jets are randomly oriented, which prompts suspicion they derive from explosions of active parts of the photosphere. The spectroscopy proves the jets move away from the star at different speeds, confirming multiple events and directions as with coronal mass ejections.[30] Multiple asymmetric mass loss events and the ejection of the outermost material are deduced to have occurred within the last 500 to 1,000 years, while that of a knot near the star would be less than 100 years. The mass loss is due to strong convection in the tenuous outer layers of the star, associated with magnetic fields. Ejections are analogous to—but much larger than—coronal ejections of the Sun.[9][30][31]
In 1976, Lada and Reid[c] published observations of the bright-rimmed molecular cloud Sh2-310, which is 15″ east of the star. At its edge bordered by the bright rim, an abrupt decrease in the CO emission and an increase in brightness of the 12 CO emission were observed, indicating possible destruction of molecular material and enhanced heating at the cloud-rim interface, respectively. They assumed the distance of the cloud is approximately equal to that of the stars, which are members of the open clusterNGC 2362, that ionize the rim. NGC 2362 could be anywhere in the ranges of 1.5±0.5 kiloparsecs (kpc) or 4,890±1,630 light-years (ly) away as determined from its color-magnitude diagram.[32] This star is projected onto the tip of the cloud rim, strongly suggesting its association. Furthermore, all the vectors of velocity of Sh2-310 are very close to those of the star. There is thus a near-certain physical association of the star with Sh2-310 and with NGC 2362 in all standard models.[33]
Melnik and others later prefer a range centred on 1.2 kiloparsecs (about 3,900 light-years).[34]
Distances can be calculated by measuring the change in position against very distant background objects as the telescope orbits the Sun. However, this star has a small parallax due to its distance, and standard visual observations have a margin of error too large for a hypergiant star with an extended CSE to be useful, for example, the Hipparcos Catalogue of 1997 gives a purely notional parallax of 1.78±3.54 milliarcseconds (mas), in which the "central" figure equates to 562 pc (1,832 ly).[35] Parallax can be measured to high accuracy from the observation of masers using a long baseline interferometry. In 2008, such observations of H 2O masers using VERA interferometry from the National Astronomical Observatory of Japan gave a parallax of 0.88±0.08 mas, corresponding to a distance of 1.14+0.11 −0.09 kpc (about 3,720+360 −300 ly).[36] In 2012, observations of SiO masers using very-long-baseline interferometry (VLBI) from Very Long Baseline Array (VLBA) independently derived a parallax of 0.83±0.08 mas, corresponding to a distance of 1.20+0.13 −0.10 kpc (about 3,910+423 −326 ly).[10] These imply the cloud (Sh2-310) is less remote than thought or that the star is a foreground object.[16]
The Gaia mission provides highly constrained parallaxes to some objects, but the data release 2 value of −5.92±0.83 mas for this star is not meaningful.[37]
The spectrum of VY Canis Majoris is that of a high-luminosity M-class star. The hydrogen lines, however, have P Cygni profiles fit for luminous blue variables. The spectrum is dominated by TiO bands whose strengths suggest a classification of M5. The H-alpha (Hα) line is not visible yet and there are unusual emission lines of neutral elements such as sodium and calcium. The luminosity class as determined from different spectral features varies from bright giant (II) to bright supergiant (Ia), with a compromise being given: as M5eIbp. Old classifications were confused by the interpretation of surrounding nebulosity as companion stars.[41]
The present spectral classification system is inadequate to this star's complexities. The class depends on which of its complex spectral features are stressed. Further, key facets vary over time as to this star. It is cooler and thus redder than M2, and is usually classified between M3 and M5. A class as extreme as M2.5 appeared in a study of 2006.[4] The luminosity class is likewise confused and often given only as I, partly because luminosity classes are poorly defined in the red and infrared portions of the spectrum. One study though, gives a luminosity class of Ia+ which means a hypergiant or extremely luminous supergiant.[42]
A very large and luminous star, VY CMa is among the most extreme stars in the Milky Way and has an effective temperature below 4,000 K (3,730 °C; 6,740 °F). It occupies the upper-right hand corner of the HR diagram although its exact luminosity and temperature are uncertain. Most of the properties of the star depend directly on its distance.
Luminosity
The bolometric luminosity (Lbol) of VY CMa can be calculated from spectral energy distribution or bolometric flux, which can be determined from photometry in several visible and infraredbands. Earlier calculations of the luminosity based on an assumed distance of 1.5 kpc (4,900 ly) gave luminosities between 200,000 and 560,000 times the Sun's luminosity (L☉).[14][32][43] This is considerably very close or beyond the empirical Humphreys–Davidson limit. One study gave nearly one million L☉ at a distance of 2.1 kpc (6,800 ly).[44] In 2006 a luminosity of 430,000 L☉ was calculated by integrating the total fluxes over the entire nebula, since most of the radiation coming from the star is reprocessed by the dust in the surrounding cloud.[31] More recent estimates of the luminosity extrapolate values below 350,000 L☉ based on distances below 1.2 kpc.[3][36][45]
Most of the output of VY CMa is emitted as infrared radiation, with a maximum emission at 5–10 μm, which is in part caused by reprocessing of the radiation by the circumstellar nebula.[9][31] Many older luminosity estimates are consistent with current ones if they are re-scaled to the distance of 1.2 kpc.[36] Despite being one of the most luminous stars in the Milky Way, much of the visible light of VY CMa is absorbed by the circumstellar envelope, so the star needs a telescope to be observed. Removing its envelope, the star would be one for the naked eye.[24]
Mass
Since this star has no companion star, its mass cannot be measured directly through gravitational interactions. Comparison of the effective temperature and bolometric luminosity compared to evolutionary tracks for massive stars suggest its initial mass was 25±10 M☉ for a rotating star but current mass 15 M☉—or 32 M☉ at first if non-rotating falling to present-day 19 M☉,[3] and an age of 8.2 million years (Myr).[10] Older studies have found much higher initial masses (thus also higher current masses) or a progenitor mass of 40–60 M☉ based on old luminosity estimates.[14][46]
Mass loss
Image taken by the ESO's Very Large Telescope showing the asymmetric nebula around VY CMa using SPHERE instrument. The star itself is hidden behind a dark disk. Crosses are artifacts (lens effects) due to the characteristics of the instrument.
VY CMa has a strong stellar wind and is losing much material due to its high luminosity and quite low surface gravity. It has an average mass loss rate of 6×10−4M☉ per year, among the highest known and unusually high even for a red supergiant, as evidenced by its extensive envelope.[47][39] It is thus an exponent for the understanding of high-mass loss episodes near the end of massive star evolution.[48] The mass loss rate probably exceeded 10−3M☉/yr during the greatest mass loss events.[47]
The star has produced large, probably convection-driven, mass-loss events 70, 120, 200, and 250 years ago. The clump shed by the star between 1985 and 1995 is the source of its hydroxyl maser emission.[49]
Temperature
The effective temperature of this star is uncertain. Some signature changes in its spectrum correspond to temperature variations. Early estimates of the mean temperature assumed values below 3,000 K based on a spectral class of M5.[43][44] In 2006, its temperature was calculated to be as high as 3,650±25 K, corresponding to a spectral class of M2.5,[4] yet this star is usually considered as an M4 to M5 star. Adopting the latter classes with the temperature scale proposed by Emily Levesque gives a range of between 3,450 and 3,535 K.[50]
Size
Right to left: VY Canis Majoris compared to Betelgeuse, Rho Cassiopeiae, the Pistol Star, and the Sun (too small to be visible in this thumbnail). The orbits of Jupiter and Neptune are also shown.
The calculation of the radius of VY CMa is complicated by the extensive circumstellar envelope of the star. VY CMa is also a pulsating star, so its size changes with time. Earlier direct measurements of the radius at infrared (K-band = 2.2 µm) wavelength gave an angular diameter of 18.7±0.5 mas, corresponding to radii above 3,000 R☉ (2.1×109 km; 14 au; 1.3×109 mi) at an assumed distance of 1.5 kpc, considerably larger than expected for any red supergiant or red hypergiant.[43] However, this is probably larger than the actual size of the underlying star and the angular diameter estimate appears exceedingly large due to interference by the circumstellar envelope.[31][9][3] In 2006–2007 radii of 1,800–2,100 R☉ have been derived from the estimated luminosity of 430,000 L☉ and temperatures of 3,450–3,535 K.[31][9]
On 6 and 7 March 2011, VY CMa was observed at near-infrared wavelengths using interferometry at the Very Large Telescope. The size of the star was calculated using the Rosseland Radius, the location at which the optical depth is 2⁄3,[51] with two modern distances of 1.14+0.11 −0.09 and 1.20+0.13 −0.10 kpc.[36][10] Its angular diameter was directly measured at 11.3±0.3 mas, which corresponds to a radius of 1,420±120 R☉ at a distance of 1.17+0.08 −0.07 kpc. The high spectral resolution of these observations allowed the effects of contamination by circumstellar layers to be minimised. An effective temperature of 3,490±90 K, corresponding to a spectral class of M4, was then derived from the radius and a luminosity of 270,000±40,000 L☉ which is based on the distance and a measured flux of (6.3±0.3)×10−13 W/cm2.[3] In late 2013, a radius of 2,069 R☉ was determined, based on a rather cool adopted temperature of 2,800 K and a luminosity of 237,000 L☉.[52]
Most radius estimates of the VY CMa are considered as the size for the optical photosphere while the size of the star for the radio photosphere is calculated to be twice that of the size of the star for the optical photosphere.[5] Despite the mass and very large size (though some estimates give smaller sizes), VY CMa has an average density of 5.33 to 8.38 mg/m3 (0.00000533 to 0.00000838 kg/m3), it is over 100,000 times less dense than Earth's atmosphere at sea level (1.2 kg/m3).
VY Canis Majoris has been known to be an extreme object since the middle of the 20th century, although its true nature was uncertain.[41][53] In the late 20th century, it was accepted that it was a post-main sequence red supergiant. Its angular diameter had been measured and found to be significantly different depending on the observed wavelength. The first meaningful estimates of its properties showed a very large star.[54][55]
Early direct measurements of the radius at infrared (K-band = 2.2 µm) wavelength gave an angular diameter of 18.7±0.5 mas, corresponding to radii above 3,000 R☉ (2.1×109 km; 14 au; 1.3×109 mi) at a still very plausible distance of 1.5 kiloparsecs; a radius dwarfing other known red hypergiants.[43] However, this is probably larger than the actual size of the underlying star—this angular diameter estimate is heightened from interference by the envelope.[3][9][31] In 2006–07, radius between 1,800–2,100 R☉ has been derived from the preferring luminosity of 430,000 L☉ and the still-preferred temperature range of 3,450–3,535 kelvin.[9][31]
In contrast to prevailing opinion, a 2006 study, ignoring the effects of the circumstellar envelope in the observed flux of the star, derived a luminosity of 60,000 L☉, suggesting an initial mass of 15 M☉ and radius of 600 R☉ based on an assumed effective temperature of 3,650 K and distance of 1.5 kpc. On this basis they considered VY CMa and another notable extreme cool hypergiant star, NML Cygni, as normal early-type red supergiants.[4][56] They assert that earlier very high luminosities of 500,000 L☉ and very large radii of 2,800–3,230 R☉[14][57] (or even 4,000 R☉[20]) were based on effective temperatures below 3,000 K that were unreasonably low.[4]
Almost immediately another paper published a size estimate of 1,800–2,100 R☉ and concluded that VY CMa is a true hypergiant. This uses the later well-reviewed effective temperature 3450–3535 kelvin, and a luminosity of 430,000 L☉ based on SED integration and a distance of 1.5 kpc.[31]
In 2011,[d] the star was studied at near-infrared wavelengths using interferometry at the Very Large Telescope. The size of the star was published at its Rosseland Radius, outside of which optical depth falls below 2⁄3,[51] given the mean of two most modern, similar but distinct distances.[e][10][36] Its angular diameter was directly measured at 11.3±0.3 mas, thus radius of 1,420±120 R☉ given a distance of 1.17+0.08 −0.07 kpc. The high spectral resolution of these observations allowed the effects of contamination by circumstellar layers to be minimised. An effective temperature of 3,490±90 K, corresponding to a spectral class of M4, was then derived from the radius and a luminosity of 270,000±40,000 L☉ which is based on the distance and a measured flux of (6.3±0.3)×10−13 W/cm2.[3]
Most such radius estimates are considered as the size for the mean limit of the optical photosphere while the size of the star for the radio photosphere is calculated to be twice that.[5] Despite the mass and very large size (though some estimates give smaller sizes), VY CMa has an average density of 5.33 to 8.38 mg/m3 (0.00000533 to 0.00000838 kg/m3). It is over 100,000 times less dense than Earth's atmosphere at sea level (1.2 kg/m3).
In 2012 the size was calculated more accurately to be somewhat lower, for example 1,420 R☉,[3] which leaves larger sizes published and in-date for other galactic and extragalactic red supergiants (and hypergiants) such as Westerlund 1 W26 and WOH G64. Despite this, VY Canis Majoris is still often described as the largest known star, sometimes with caveats to account for the highly uncertain sizes of all these stars.[58][f] A 2013-made estimate based on the Wittkowski radius and the Monnier radius put mean size at 2,000 R☉,[27] and later that year, Matsuura and others put forward a competing method of finding radius within the envelope, putting the star at 2,069 R☉, based on a cool-end of estimates adopted temperature of 2,800 K and a luminosity of 237,000 L☉.[52] However, these values are not consistent with its spectral types, leaving the 2012 values in better match.
Surroundings
WFPC2/HST image showing the asymmetric nebula surrounding VY CMa, which is the central star
VY Canis Majoris is surrounded by an extensive and dense asymmetric red reflection nebula, with a total ejected mass of 0.2-0.4 M☉ and a temperature of 800 kelvin, based on a DUSTY model atmosphere that has been formed by material expelled from its central star.[14][47] The inner shell figures as 0.12 ″ across, corresponding to 140 AU (0.0022 ly) for a star 1,200 parsecs away, whereas that of the outer one is at 10″, corresponding to 12,000 AU (0.19 ly).[47] This nebula is so bright that it was discovered in a dry night sky in 1917 with an 18 cm telescope, and its condensations were once regarded as companion stars.[20] It has been extensively studied with the aid of the Hubble Space Telescope (HST), showing that the nebula has a complex structure that includes filaments and arcs, which were caused by past eruptions; the structure is akin to that around post-red supergiant (Post-RSG) or yellow hypergiant (YHG) IRC +10420. The similarity has led at least two professional articles to propose a model that the star might evolve blueward on the Hertzsprung–Russell diagram (HR diagram) to become a yellow hypergiant, then a luminous blue variable (LBV), and finally a Wolf–Rayet star (WR star).[14][20]
Evolution
VY Canis Majoris is a highly evolved star yet less than 10 million years (Myr) old. Some old writings envisaged the star as a very young protostar or a massive pre-main-sequence star with an age of only 1 Myr and typically a circumstellar disk.[15] It has probably evolved from a hot, dense O9main sequence star of 5–20 R☉ (solar radii).[30][32][59] The star has evolved rapidly because of its high mass. The time spent to the red hypergiant phase is estimated to be between 100,000 and 500,000 years, and thus VY CMa most likely left its main sequence phase more than a million years ago.[10][30]
The future evolution of VY CMa is uncertain, but like the most cool supergiants, the star will certainly explode as a supernova. It has begun to fuse helium into carbon en masse.[g] Like Betelgeuse, it is losing mass and is expected to explode as a supernova within the next 100,000 years — it will probably revert to a higher temperature beforehand.[3][58][60] The star is very unstable, having a prodigious mass loss such as in ejections.
VY Canis Majoris is a candidate for a star in a second red supergiant phase, but this is mostly speculative and unconfirmed.[61]
From this star CO emission is coincident with the bright KI shell in its asymmetric nebula.
or less likely, a type Ib supernova, but it is unlikely that would be as luminous as SN 2006tf or SN 2006gy.
The explosion could be associated with gamma-ray bursts (GRB), and it will produce a shock wave of a speed of a few thousand kilometers per second that could hit the surrounding envelope of material, causing strong emission for many years after the explosion. For a star so large, the remnant would be probably a black hole rather than a neutron star.[60]
Notes
^The given stellar metallicity is given as the solar metallicity ([Fe/H] = approx. +0.0 dex).
^1.14+0.11 −0.09 and 1.20+0.13 −0.10 kpc, see above at Distance
^ Alcolea et al 2013 refer to VY CMa as having the highest radius "among well-characterised stars in our galaxy", referring to the Wittkowski et al. 2012 value of 1,420 R☉ which is based on the distances from Choi et al. 2008 and Zhang et al. 2012 plus an angular diameter. Several red supergiants (or hypergiants) are possibly larger, although they could have less accurate radius estimates.
^ abcdefDucati, J. R (2002). "VizieR Online Data Catalog: Catalogue of Stellar Photometry in Johnson's 11-color system". VizieR On-line Data Catalog. Bibcode:2002yCat.2237....0D.
^ abSerkowski, K (1969). "Large Optical Polarization of the OH Emission Source VY Canis Majoris". The Astrophysical Journal. 156: L139. Bibcode:1969ApJ...156L.139S. doi:10.1086/180366.
^ abcRobinson, L. J. (1971). "Three Somewhat Overlooked Facets of VY Canis Majoris". Information Bulletin on Variable Stars. 599: 1. Bibcode:1971IBVS..599....1R.
^ abcdeWittkowski, M.; Langer, N.; Weigelt, G. (2004). "Diffraction-limited speckle-masking interferometry of the red supergiant VY CMa". Astronomy and Astrophysics. 340 (2004): 77–87. arXiv:astro-ph/9811280. Bibcode:1998A&A...340L..39W.
^ abKamiński, T; Gottlieb, C. A; Menten, K. M; Patel, N. A; Young, K. H; Brünken, S; Müller, H. S. P; McCarthy, M. C; Winters, J. M; Decin, L (2013). "Pure rotational spectra of TiO and TiO2 in VY Canis Majoris". Astronomy and Astrophysics. 551 (2013): A113. arXiv:1301.4344. Bibcode:2013A&A...551A.113K. doi:10.1051/0004-6361/201220290. S2CID59038056.
^Lada, C. J.; Reid, M. (1976). "The Discovery of a Molecular Cloud Associated with VY CMa". Bulletin of the American Astronomical Society. 8: 322. Bibcode:1976BAAS....8R.322L.
^Perryman, M. A. C.; Lindegren, L.; Kovalevsky, J.; Hoeg, E.; Bastian, U.; Bernacca, P. L.; Crézé, M.; Donati, F.; Grenon, M.; Grewing, M.; Van Leeuwen, F.; Van Der Marel, H.; Mignard, F.; Murray, C. A.; Le Poole, R. S.; Schrijver, H.; Turon, C.; Arenou, F.; Froeschlé, M.; Petersen, C. S. (1997). "The HIPPARCOS Catalogue". Astronomy and Astrophysics. 323: L49. Bibcode:1997A&A...323L..49P.
^ abcdeChoi, Y. K.; Hirota, Tomoya; Honma, Mareki; Kobayashi, Hideyuki; Bushimata, Takeshi; Imai, Hiroshi; Iwadate, Kenzaburo; Jike, Takaaki; Kameno, Seiji; Kameya, O.; Kamohara, R.; Kan-Ya, Y.; Kawaguchi, N.; Kijima, M.; Kim, M. K.; Kuji, S.; Kurayama, T.; Manabe, S.; Maruyama, K.; Matsui, M.; Matsumoto, N.; Miyaji, T.; Nagayama, T.; Nakagawa, A.; Nakamura, K.; Oh, C. S.; Omodaka, T.; Oyama, T.; Sakai, S.; et al. (2008). "Distance to VY Canis Majoris with VERA". Publications of the Astronomical Society of Japan. 60 (5): 1007. arXiv:0808.0641. Bibcode:2008PASJ...60.1007C. doi:10.1093/pasj/60.5.1007. S2CID15042252.
^ abHumphreys, E. M. L; Immer, K; Gray, M. D; De Beck, E; Vlemmings, W. H. T; Baudry, A; Richards, A. M. S; Wittkowski, M; Torstensson, K; De Breuck, C; Moller, P; Etoka, S; Olberg, M (2017). "Simultaneous 183 GHz H2O Maser and SiO Observations Towards Evolved Stars Using APEX SEPIA Band 5". Astronomy & Astrophysics. 603: A77. arXiv:1704.02133. Bibcode:2017A&A...603A..77H. doi:10.1051/0004-6361/201730718. S2CID55162530.
^Kastner, Joel (1996). "FOC Imaging of the Dusty Envelopes of Mass-Losing Supergiants". HST Proposal: 6416. Bibcode:1996hst..prop.6416K.
^Skiff, B. A. (2014). "VizieR Online Data Catalog: Catalogue of Stellar Spectral Classifications (Skiff, 2009-2016)". VizieR On-line Data Catalog: B/Mk. Originally Published in: Lowell Observatory (October 2014). 1: B/mk. Bibcode:2014yCat....1.2023S.
^ abcdMonnier, J. D; Millan-Gabet, R; Tuthill, P. G; Traub, W. A; Carleton, N. P; Coudé Du Foresto, V; Danchi, W. C; Lacasse, M. G; Morel, S; Perrin, G; Porro, I. L; Schloerb, F. P; Townes, C. H (2004). "High-Resolution Imaging of Dust Shells by Using Keck Aperture Masking and the IOTA Interferometer". The Astrophysical Journal. 605 (1): 436–461. arXiv:astro-ph/0401363. Bibcode:2004ApJ...605..436M. doi:10.1086/382218. S2CID7851916.
^ abLe Sidaner, P; Le Bertre, T (1996). "Optical and infrared observations of 27 oxygen-rich stars. Modelling of the circumstellar dust shells". Astronomy and Astrophysics. 314: 896. Bibcode:1996A&A...314..896L.
^Levesque, Emily M.; Massey, Philip; Olsen, K. A. G.; Plez, Bertrand; Josselin, Eric; Maeder, Andre; Meynet, Georges (2005). "The Effective Temperature Scale of Galactic Red Supergiants: Cool, but Not as Cool as We Thought". The Astrophysical Journal. 628 (2): 973–985. arXiv:astro-ph/0504337. Bibcode:2005ApJ...628..973L. doi:10.1086/430901. S2CID15109583.
^ abMatsuura, Mikako; Yates, J. A.; Barlow, M. J.; Swinyard, B. M.; Royer, P.; Cernicharo, J.; Decin, L.; Wesson, R.; Polehampton, E. T.; Blommaert, J. A. D. L.; Groenewegen, M. A. T. (30 October 2013). "Herschel SPIRE and PACS observations of the red supergiant VY CMa: analysis of the molecular line spectra". Monthly Notices of the Royal Astronomical Society. 437 (1): 532–546. arXiv:1310.2947. Bibcode:2014MNRAS.437..532M. doi:10.1093/mnras/stt1906. ISSN0035-8711. S2CID53393704.
^Massey, Philip; Levesque, Emily M; Plez, Bertrand; Olsen, Knut A. G; Bresolin, F; Crowther, P. A; Puls, J (2008). "The Physical Properties of Red Supergiants: Comparing Theory and Observations". Massive Stars as Cosmic Engines. 250: 97–110. arXiv:0801.1806. Bibcode:2008IAUS..250...97M. doi:10.1017/S1743921308020383. S2CID15766762.
^ abAlcolea, J; Bujarrabal, V; Planesas, P; Teyssier, D; Cernicharo, J; De Beck, E; Decin, L; Dominik, C; Justtanont, K; De Koter, A; Marston, A. P; Melnick, G; Menten, K. M; Neufeld, D. A; Olofsson, H; Schmidt, M; Schöier, F. L; Szczerba, R; Waters, L. B. F. M (2013). "HIFISTARSHerschel/HIFI observations of VY Canis Majoris. Molecular-line inventory of the envelope around the largest known star". Astronomy & Astrophysics. 559: A93. arXiv:1310.2400. Bibcode:2013A&A...559A..93A. doi:10.1051/0004-6361/201321683. S2CID55758451.
^Wallerstein, G (1978). "An interpretation of the apparent orbit of VY CMa AB: The rotating holey dust cloud hypothesis". The Observatory. 98: 224. Bibcode:1978Obs....98..224W.
