Peripatus Home Page pix1Black.gif (807 bytes) Astronomy and Astrophysics >> Luminous Blue VariablesUpdated: 28-May-2020 

Luminous Blue Variables


ROUGH NOTES ONLY - NOT A FINISHED PAGE: The defining characteristics of luminous blue variables are presented and interpreted in an evolutionary context. Some occurences of these extremely rare objects are listed with their known physical properties.

Keywords: luminous blue variable, LBV, P Cygni, S Doradus, Hubble-Sandage


LBVs are massive, intrinsically bright stars which display different scales of light and colour variability, ranging from rapid micro-variations to rare outbreaks of catastrophic mass loss. They represent a very short-lived (perhaps as little as 40,000 years) strongly mass-losing phase in the evolution of massive stars, during which they undergo deep erosion of the outer layers before entering the Wolf-Rayet phase.

They include P Cygni stars, S Doradus stars, and Hubble-Sandage variables.

The LBV class was first defined by Conti (1984).



LBV spectra typically exhibit prominent emission lines of H, HeI, FeII and [FeII], often with P Cygni profiles [→ sidebar]. These emission lines are much stronger than normal for the temperature and luminosity of the stars.

The spectra, luminosity and temperature are highly variable. During quiet periods of minimum brightness (so-called ‘quiescence’) they appear as blue, B-type supergiants with temperatures greater than around 15 – 20,000 K showing H and HeI emission. During periods of maximum brightness they appear as later A- or F-type supergiants with temperatures near 8000 K and FeII and [FeII] emission reaches maximum. It is widely believed that the bolometric magnitudes, i.e. total radiation, remain roughly constant throughout these cycles and that the apparent cooling is due to absorption by surrounding shells of ejected matter which re-radiate the stars’ visible and blueward radiation at longer wavelengths (Parker et al. 1993).

P Cygni spectra are characterised by emission lines flanked on their violet edges by absorptions and are produced by an expanding envelope of gas or stellar wind.

Other than that which is “blowing” straight towards or away from us, the gas streaming away from the star is travelling more or less perpendicular to our line of sight, producing emission lines which are not Doppler shifted. The gas which is travelling straight towards us, however, lies directly in front of the star, so producing an absorption feature in the stellar continuum, which is Doppler shifted to shorter wavelengths and appears on the violet edge of the emission lines. The gas streaming directly away from us on the far side of the star is not seen at all.

(After Kaler 1989, pp. 193-194. See also Crowther 1997, pp. 51-57.)

Mass and Luminosity

Luminosity is typically in the order of 106 L (Mbol ~ -10), which is near the observed upper luminosity limit of all stars, and their episodic outbursts may be a clue to the mechanism which brings about the limitin the first place.

From their luminosity and positions on the H-R diagram, we can categorise LBVs into three broad luminosity types:

  • The most luminous star, h Carinae, occupies a class of its own.
  • The second group contains stars which range from -11 > Mbol > -9.9 and includes most of the well-known LBVs: R127, S Dor, P Cyg and AG Car.
  • The lowest luminosity LBVs are found below the Humphreys-Davidson limit and are typified by R71.

The underlying basis for these apparently natural groupings may lie in the available range for radius variation before the star runs into whatever physical process drives the Humphreys-Davidson limit.

(After Bohannan 1989, p. 36.)

Puls et al. 2008, p. 52-53:

Reliable determinations of stellar properties in LBVs became possible only after the advent of unified model atmospheres. Even then, however, in some extreme objects like h Car, the enormous complexity of the observed ground based spectra (Hillier & Allen 1992), contaminated by multiple emission regions, hinders a proper quantitative spectroscopic analysis. In this case, stellar parameters need to be determined either from “cleaner” wavelength regions (e.g., from the millimeter continuum, Cox et al. 1995) or from “cleaner”, resolved observations (e.g., using the HST, Hillier et al. 2001, or the recent, really promising results from long baseline interferometers such as AMBER-VLTI, Weigelt et al. 2007).

“The outstanding wind density (with ~ 10-3M yr-1 as estimated for this object, Cox et al. 1995, Hillier et al. 2001) places R(Ross = 2/3) at 80% of the terminal velocity, impeding any derivation of the hydrostatic radius. Nevertheless, CNO abundances could be determined and are consistent with those found for the surrounding nebula (for a recent review, see Najarro & Hillier 2008 [2012?]).”

Variability/Mass Loss

Most characteristically, LBVs exhibit variability ranging from day to day microvariations of around 0.1 to 0.2 magnitudes, through ‘normal’ variations of 1 to 2 magnitudes on time scales of years to a few decades, to rare eruptions which can increase their brightness by more than 3 magnitudes occurring less frequently, perhaps once in a few centuries.

  1. LBVs exhibit microvariations of around 0.1 to 0.2 magnitudes on a time scale of days or a few tens of days. Wolf (1989, p. 93) notes recorded time scales for R71 of ~23.5 days between 1983 and 1985, shortening to ~14.3 days between 1986 and 1987. “[It] is easily seen that even small photometric variations are paralleled by significant spectroscopic variations in the sense that the spectrum becomes later when the star becomes visually brighter.” Van Genderen et al. 1990 reports similar findings for AG Car (P ~ 10 days) and HR Car (P ~ 20 days).
  2. During ‘normal’ (S Doradus type) eruptions, the stars may brighten by up to two magnitudes within a few months.

    At these times the stars’ atmospheres form a slowly (100 to 200 km s-1) expanding “optically thick pseudo-photosphere” (Humphreys 1989, p. 4). Mass loss increases 10 to 100 times from quiescent levels, which are near those of normal supergiants of the same temperature and luminosity, up to 10-5 to 10-4M yr-1. Both the minima and the maxima may last several years (Wolf 1989, p. 92). Superimposed over these variations are smaller oscillations of around half a magnitude, occurring on a time scale of months to a few years.

    Stothers & Chin (1995) found that the periodicity of the normal variations seems to be related to absolute bolometric magnitude as shown in the following table.

    h CarMilky Way-11.63-5 years
    AF AndM31-11.43-7
    AG CarMilky Way-10.84-14
    Var. BM33-10.25-8
    P CygMilky Way-9.9~18
    S DorLMC-9.810-30
    Var. CM33-9.819-20

  3. LBVs are perhaps most famous for their sporadic giant (‘plinian’ or ‘Eta Carinae type’) eruptions which may represent the causal mechanism behind the Humphreys-Davidson limit. During these eruptions, large amounts of mass are suddenly ejected. The most famous example is certainly the brightening of h Carinae between 1837 and 1860, when it outshone Canopus. Less well-known, P Cygni brightened from below naked-eye visibility to around third magnitude in the 1600s.

    The interval between major eruptions, during which the stars remain quiescent, is estimated in the order of a few hundreds to a few thousands of years.

    The sporadic ejections of mass and high rates of steady mass loss at other times surrounds the LBV with a circumstellar shell of dust which may even be clearly visible as in the case of the Homunculus Nebula surrounding h Car or the ring nebula surrounding AG Carinae. Over the life-time of the LBV, usually estimated as ~105 years, they may shed around 1 M or more through steady mass loss and more during major eruptions. Stahl 1987 estimates the average mass loss rate for a number of LBVs, including AG Car and R127, to be >10-4 M yr-1.

    Analyses have indicated that the ejecta are rich in nitrogen and helium, presumably arising from fusion in the core and brought to the surface by mixing or convective overshooting.

    The lifetime of LBVs can be estimated from the relative numbers of LBVs versus other stars. For example, in the LMC there are known six LBVs and 115 Wolf-Rayet stars. “If one assumes that there are ten current LBVs to account for a few unknown, the phase between core-hydrogen and core-helium burning would then be about one-twelfth of the entire W-R lifetime (5 x 105y), ~ 40,000 years” (Bohannan 1997, p.4).

“Whilst most Galactic and Magellanic Cloud LBVs have been subject to photometric monitoring, only few have been analyzed spectroscopically in sufficient detail to understand the driving mechanism of their winds. As reported above, mass-loss rates are of the order of 10-3 – 10-5 M yr-1, whilst terminal wind velocities are in the range ~100 -500 kms-1. Of course, these values vary with L and M, but there are also indications that the mass loss varies as a function of Teff when the SDor variables transit the upper HRD on timescales of years. It is this aspect that provides us an ideal laboratory for testing the theory of radiation-driven winds” (Puls et al. 2008, p. 55).



“Recently the classical view of luminous blue variables (LBVs) as a phase of the post-main sequence (MS) evolution of a single massive star has been challenged by Smith & Tombleson (2015), who proposed that LBVs are mass-gainers in binary systems with Wolf Rayet (WR) stars, which are the mass-donors. Humphreys et al. (2016) defended the accepted description of LBVs as evolved massive stars that have to quickly lose their H envelope through severe mass-loss, before to [sic] evolve as WRs. The debate is very active and the mass-loss mechanism that triggers the LBV Giant Eruptions (Humphreys & Davidson 1994) is still not understood” (Agliozzo et al. 2016, p. 69).

It is widely believed that most stars with an initial mass greater than about 40 M go through a short (up to about 105 years) LBV phase in the late stages of their evolution, after leaving the main sequence (e.g. Drissen et al. 1997). Moffat et al. 1989 (p. 229) goes even further, suggesting that “the LBV mechanism is an essential step to ‘force’ massive stars (MZAMS / 40 M ) to finally enter the Wolf-Rayet (W-R) domain”.

Stars somewhat below this mass may become variable red supergiants showing some LBV characteristics (occasionally referred to as “red LBVs,” e.g. by Moffat et al. 1989, p.234). This class includes the highly variable OH/IR stars. Note, however, that there are some supposed LBVs which have relatively low luminosity and interpreted low mass: perhaps as low as 20 to 30 M. This group includes R66, AE Andromedae, and ‘var. 15’ from M31. Their evolutionary future is unclear (Moffat et al. 1989, p.234).

The obvious place to seek candidate progenitor stars is in adjacent areas of the HR Diagram, between the H-ZAMS and the Humphreys-Davidson Limit. This area includes early type luminous supergiants “that have not yet shown any obvious LBV activity beyond quiescent microvariability.... They evolve essentially horizontally in the H-R Diagram, both before and after encountering the H-D Limit” (Moffat et al. 1989, p.229).

As noted above, detailed photometric studies reveal microvariability in “non-variable” supergiants, most pronounced in early (O and early B) type stars of high bolometric magnitude, the amplitudes of which “resemble the microvariations observed in LBVs during quiescence” (Moffat et al. 1989, p.230).


During the so-called “normal” (S Doradus type) outbursts, the objects evolvered-wards to cooler temperatures, bringing them into contact with the Humphreys-Davidson limit and stimulating accelerated mass loss. “After an LBV has lost enough mass in an eruption, the instability is temporarily relieved and the star returns to its quiescent[blue] state, presumably until the instability develops again” (Humphreys 1989, p. 7).

Stars may run into several LBV episodes. “As they careen across the H-R diagram inrepeated blue-red-blue excursions, they lose mass through constant winds and occasional outbursts. This scenario may lead an LBV to a Wolf-Rayet phase or even a supernova, which perhaps could occur either during the outburst phase as a red supergiant or the quiescent phase as a blue supergiant such as the progenitor of SN 1987A. These temperature variations should not be confused with the ‘blue loops’ that are thought to occur in the post-main-sequence evolution of massive stars due to nuclear evolution; in LBVs the variability is due to photospheric instabilities and changes” (Parker et al. 1993, p. 770; also see Bohannan 1997).

“Ofpe/WN9 stars appear also to be related to the LBV phenomenon. The best known examples of this connection are R127 and AG Carinae. R127 had been originally classified as an Ofpe/WN9 star till the ’80s when Stahl et al. (1983) observed it undergoing a typical LBV outburst. AG Carinae is a proven LBV star which assumes an Ofpe/WN9-like spectral morphology during its quiescent periods (Stahl 1987). As a consequence of these observations Ofpe/WN9 stars are now suspected to be dormant LBVs” (Pasquali et al. 1996).

“This relationship has been further explored by Smith, Crowther, & Prinja (1994) who propose the some LBVs are the extension of the WN sequence towards later spectral types, hence unifying the classes of WNL, LBV, and Ofpe/WN9 stars” (Nota et al. 1996, p. 385).

“The above properties make Ofpe/WN9 stars ideal to verify stellar evolution models and to study the mass loss phenomenon. In particular, it is interesting to compare the mass loss parameters (such as mass loss rate, stellar radiation field and wind momentum) found for Ofpe/WN9 with the ones derived for O and WR stars. In this way, it becomes possible to compare the mass loss properties during the evolution from the main sequence to the early Wolf-Rayet stage” (Pasquali et al. 1996).

Higher mass-loss rate and enhanced mixing between core and envelope are required in order to yield models compatible with the observed properties of Ofpe/WN9 stars. The emerging picture may be consistent with earlier evidence of Ofpe/WN9 stars being quiescent LBVs. This idea is further strengthened by the highly reduced surface H mass fractions of the Ofpe/WN9 stars. We derive Xs = 0.5 to 0.3, which still excludes Ofpe/WN9stars from being core He burning objects, but is almost identical to the Xs values recently measured in LBVs [REFERENCE?].

A number of Ofpe/WN9 star spectra show the presence of nebular emission lines, indicating a surrounding nebulosity though only in some cases (S119, BE 381) is there a clear case for an expanding shell associated with the star. “This would strengthen the connection between Ofpe/WN9 stars and luminous blue variables, which also are often surrounded by circumstellar nebulae generated during one or more ‘violent outbursts’ in their recent evolution history (Nota et al. 1996, p. 383).


H-depletion and N-enrichment have been noted in some LBVs (see Langer 1990 for references) so “it does seem likely that after several major LBV eruptions, one should look among luminous stars with enhanced CNO-cycle abundances. The W-R stars are the most obvious candidates, in particular the most luminous ones, of type WNL (late-type W-R stars of subtypes WN6-9)” (Moffat et al. 1989, p.229). An example supporting this conjecture is R127 in the LMC, originally recorded as an Of+WN binary until its recent eruption.

Occurrence and Examples

LBVs are among the rarest of stars; only about 32 stars in the local group of galaxies are known to be LBVs and six of these are in the LMC (Parker et al. 1993, p. 770; Bohannan 1997).


Some other examples are HR Car in the Milky Way (see Agliozzo et al. 2016), AE And and AF And in M31, and RMC127 (Agliozzo et al. 2016) and HDE 269582 in the LMC.

HR Car in the Milky Way (see Agliozzo et al. 2016)

During the quiescence, HR Car is a B2 supergiant and is not hot enough to ionize the circumstellar nebula, therefore White 2000 suggested that HR Car may have a Recently, Boffin et al. 2016 obtained a direct detection of at least one companion star for HR Car, very likely a red supergiant (RSG). The Herschel images indicate that HR Car’s optical and radio nebula is surrounded by an extended and asymmetric outer nebula of optically thin dust. This variation does not seem correlated with the star S Doradus cycle, so we proposed that it could have been induced by the periastron passage of the companion star. Since HR Car is in a binary or even multiple system, the star may suffer a torque that causes an axial precession and then a helical outflow. The stellar wind would be collimated along the rotational axis of HR Car, and gives origin to the Ha and free-free emission. The ionization of the ejecta has to occur thanks to a third companion, a spectral type BO V star, as suggested by White 2000.

paraphrase and add fig. 3

RMC127 in the LMC (see Agliozzo et al. 2016)

Davies et al. 2005, found a high-degree of polarization associated with the central star, suggesting an aspherical mass-loss, possibly a disk in the equatorial plane of the star. Later the HST provided the highest resolution image of the Ha nebula (Weis 2003). Weis 2003 found that the nebula deviates from a spherical symmetry, with a bipolarity in the North-South direction, forming two Caps, perpendicular to two Rims in the East-West direction. This configuration resembles the nebular features reproduced by Nota et al. 1995 by simulating a fast outflow in a preexisting dense medium. broad and shallow photospheric absorption line Si IV ?4088, as shown in Fig. 4, which suggests that RMC127 is a fast rotator. From P Cyg profiles of He I lines with the absorption component saturated we could also determine the terminal velocity of the wind (148 ± 14kms-1 ). Very likely the IR emission detected from the space telescopes arises from extended optically thin dust, rather than a compact shell close to the star. stellar wind and optical nebula by the previously mentioned authors make all the models based on spherical symmetry unsuitable for RMC127. Therefore, we modeled the central object emission with the collimated stellar wind model formulated by Reynolds 1986. We found that, in this scenario, the mass-loss rate of the outflow would be ~ 8 - 9 10-6 M yr-1 , a factor of 3 less than the spherical case.The precession needs a torque from a companion star to occur, similarly to HR Car. However, the binarity nature of RMC127 has not been yet demonstrated.

Paraphrase and add fig. 6

Another candidate LBV, R81 in the LMC, is an eclipsing binary. (For a discussion of binary scenarios, see Gallagher 1989.)


Agliozzo, C.; Trigilio, C.; Buemi, C.; Leto, P.; Umana, G.; Pignata, G.; Phillips, N.M.; Nikutta, R.; Prieto, J.L. 2016: The mass-loss before the end: two luminous blue variables with a collimated stellar wind. In Eldridge, J.J.; Bray, J.C.; McClelland, L.A.S.; Xiao, L. (ed.) 2016: The Lives and Death-Throes of Massive Stars. Proceedings IAU Symposium 329 : 69-73.

Bohannan, B. 1989: The distribution of types of luminous blue variables. In Davidson, K.; Moffat, A.F.J.; Lamers, H.J.G.L.M. (ed.) 1989: Physics of Luminous Blue Variables. Kluwer Academic Publishers : 35-43.

— 1997: The definition of luminous blue variables. In Nota, A.; Lamers, H.J.G.L.M. (ed.) 1997: Luminous Blue Variables: Massive Stars in Transition. Astronomical Society of the Pacific Conference Series 120 : 3-7.

Conti, P.S. 1984: Basic observational constraints on the evolution of massive stars. In Maeder, A.; Renzini, A. (ed.) 1984: Observational tests of stellar evolution theory. IAU Symposium 105 : 233-254.

Cox P.; Mezger P.G.; Sievers A.; Najarro F.; Bronfman L.; Kreysa E.; Haslam G. 1995: Millimeter emission of eta Carinae and its surroundings. Astronomy & Astrophysics 297: 168-174.

Crowther, P.A. 1997: Physical properties of LBVs and related objects. In Nota, A.; Lamers, H.J.G.L.M. (ed.) 1997: Luminous Blue Variables: Massive Stars in Transition. Astronomical Society of the Pacific Conference Series 120 : 51-57.

Drissen, L.; Crowther, P.A.; Smith, L.J.; Robert, C.; Roy, J.-R.; Hillier, D.J. 1997: Physical parameters of erupting Luminous Blue Variables: NGC 2363-V1 caught in the act. The Astrophysical Journal 546: 484-495.

Gallagher, J.S. 1989: Close binary models for Luminous Blue Variable stars. In Davidson, K.; Moffat, A.F.J.; Lamers, H.J.G.L.M. (ed.) 1989: Physics of Luminous Blue Variables. Kluwer Academic Publishers : 185-194.

Hillier, D.J.; Allen, D.A. 1992: A spectroscopic investigation of Eta Carinae and the Homunculus Nebula. I - Overview of the spectra. A&A 262: 153-170.

Hillier, D.J.; Davidson, K.; Ishibashi, K.; Gull, T. 2001: On the Nature of the Central Source in η Carinae. ApJ 553: 837-860.

Humphreys, R.M. 1989: What are LBVs? - Their characteristics and role in the upper H-R diagram. In Davidson, K.; Moffat, A.F.J.; Lamers, H.J.G.L.M. (ed.) 1989: Physics of Luminous Blue Variables. Kluwer Academic Publishers : 3-12.

Humphreys, R.M.; Davidson, K. 1994: The Luminous Blue Variables: Astrophysical geysers. PASP 106 (704): 1025-1051.

Humphreys, R.M.; Weis, K.; Davidson, K.; Gordon, M.S. 2016: On the social traits of Luminous Blue Variables. The Astrophysical Journal 825 (64): 1-15.

Kaler, J.B. 1989: Stars and their spectra: an introduction to the spectral sequence. Cambridge: 1-300.

Langer, N. 1990: Models of Wolf-Rayet Stars. In Garmany, C.D. 1990: Properties of hot luminous stars. Proceedings of the First Boulder-Munich Workshop, Boulder, CO, Aug. 6-11, 1988. Astronomical Society of the Pacific Conference Series 7 : 328-339.

Moffat, A.F.J.; Drissen, L.; Robert, C. 1989: Observational connections between LBVs and other stars, with emphasis on Wolf-Rayet stars. In Davidson, K.; Moffat, A.F.J.; Lamers, H.J.G.L.M. (ed.) 1989: Physics of Luminous Blue Variables. Kluwer Academic Publishers : 229-240.

Najarro, F.; Hillier, D. 2008: Reference not found; possibly published later as Najarro & Hillier 2012. .

Najarro, F.P.; Hillier, D.J. 2012: Chapter 4. The winds of Eta Carinae and other very luminous stars. In Davidson, K.; Humphreys, R.M. (ed) 2012: Eta Carinae and the Supernova Impostors. Springer, Astrophysics and Space Science Library book series 384 : 67-93.

Nota, A.; Pasquali, A.; Drissen, L.; Leitherer, C.; Robert, C.; Moffat, Anthony F. J.; Schmutz, W. 1996: O stars in transition. I. Optical spectroscopy of Ofpe/WN9 and related stars. Astrophysical Journal Supplement 102: 383-410.

Parker, J.W.; Clayton, G.C.; Winge, C.; Conti, P.S. 1993: A new luminous blue variable: R 143 in 30 Doradus. Astrophysical Journal 409: 770-775.

Pasquali, A.; Schmutz, W.; Leitherer, C.; Nota, A.; Hubeny, I.; Langer, N.; Drissen, L.; Robert, C. 1996: Fundamental properties of Ofpe/WN9 stars from ultraviolet HST spectra. Science with the HST 2: 386-392.

Puls, J.; Vink, J.S.; Najarro, F. 2008: Mass loss from hot massive stars. Astronomy and Astrophysics Review 16: 209-325.

Smith, L. J.; Crowther, P. A.; Prinja, R. K. 1994: A study of the luminous blue variable candidate He 3-519 and its surrounding nebula. Astronomy and Astrophysics 281: 833-854.

Smith, N.; Tombleson, R. 2015: Luminous blue variables are antisocial: their isolation implies that they are kicked mass gainers in binary evolution. MNRAS 447: 598-617.

Stahl, O. 1987: Direct imagery of circumstellar shells around Ofpe/WN9 stars in the galaxy and in the LMC. Astronomy & Astrophysics 182: 229-236.

Stahl, O.; Wolf, B.; Klare, G.; Cassatella, A.; Krautter, J.; Persi, P.; Ferrari-Toniolo, M. 1983: R 127: an S Dor type variable intermediate between Of and WN. Astronomy & Astrophysics 127: 49-62.

Stothers, R.B.; Chin, C. 1995: A period-luminosity relation for the slow variation of luminous blue variables. Astrophysical Journal 451: L61-L64.

van Genderen, A.M.; Thé, P.S.; Heemskerk, M.; Heynderickx, D.; van Kampen, E.; Kraakman, H.; Larsen, I.; Remijn, L.; Wanders, I.; van Weeren, N. 1990: Light variations of massive stars (α Cygni variables). XI. The behaviour of the S Doradus type stars AG Carinae and HR Carinae near minimum light. Astronomy & Astrophysics Supplement Series 82: 189-200.

Weigelt, G.; Kraus, S.; Driebe, T.; Petrov, R.G.; Hofmann, K.H.; Millour, F.; Chesneau, O.; Schertl, D.; Malbet, F.; Hillier, J.D.; and 93 coauthors 2007: Near-infrared interferometry of η Carinae with spectral resolutions of 1 500 and 12 000 using AMBER/VLTI. Astronomy & Astrophysics 464: 87-106.

Wolf, B. 1989: "Normal" LBV Eruptions a la S Doradus. In Davidson, K.; Moffat, A.F.J.; Lamers, H.J.G.L.M. (ed.) 1989: Physics of Luminous Blue Variables. Kluwer Academic Publishers : 91-99.

 Peripatus Home Page pix1Black.gif (807 bytes) Astronomy and Astrophysics >> Luminous Blue Variables

Hits counted from 4 Apr 2020: PHP Hits Count