A revised historical light curve of Eta Carinae and the timing of close periastron encounters

The historical light curve of the 19th century 'Great Eruption' of η Carinae provides a striking record of the violent instabilities encountered by massive stars. In this paper, we report and analyse newly uncovered historical estimates of
of 11
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Mon. Not. R. Astron. Soc.  415,  2009–2019 (2011) doi:10.1111/j.1365-2966.2011.18993.x A revised historical light curve of Eta Carinae and the timing of closeperiastron encounters Nathan Smith 1  and David J. Frew 2 1 Steward Observatory, University of Arizona, 933 North Cherry Avenue, Tucson, AZ 85721, USA 2  Department of Physics and Astronomy, Macquarie University, North Ryde, NSW 2109, Australia Accepted 2011 April 28. Received 2011 April 19; in srcinal form 2010 October 13 ABSTRACT The historical light curve of the 19th century ‘Great Eruption’ of  η  Carinae provides a strikingrecord of the violent instabilities encountered by massive stars. In this paper, we reportand analyse newly uncovered historical estimates of the visual brightness of   η  Car during itseruption,andwecorrectsomemistakesintheoriginalrecord.Therevisedhistoricallightcurvelooks substantially different from previous accounts; it shows two brief precursor eruptions in1838 and 1843 that resemble modern supernova impostors, while the final brightening in 1844December marks the time when  η  Car reached its peak brightness. We consider the timingof brightening events as they pertain to the binary system in  η  Car. (1) The brief 1838 and1843 events rosetopeak brightness withinweeks ofperiastronpassages ifthepre-1845 orbitalperiod was ∼ 5 per cent shorter than that at present due to the mass-loss of the eruption. Eachevent lasted only  ∼ 100d. (2) The main brightening at the end of 1844 has no conceivableassociation with periastron, beginning suddenly more than 1.5yr  after   periastron. It lasted ∼ 10yr, with no obvious influence of periastron encounters during that time. (3) The 1890eruption  began  to brighten at periastron, but took over 1yr to reach maximum brightnessand remained there for almost 10yr. A second periastron passage mid-way through the 1890eruption had no visible effect. While the evidence for a link between periastron encountersandthetwobriefprecursorevents iscompelling, thedifferences between thethreecasesabovemake it difficult to explain all three phenomena with the same mechanism. Key words:  instabilities – binaries: close – stars: individual: Eta Carinae – stars: massive –stars: mass-loss. 1 INTRODUCTION Among massive stars, the enigmatic object  η  Carinae is simulta-neously our most scrutinized case study and still one of the mostmysterious (Davidson & Humphreys 1997). Its nebula providesproof that massive stars can eject more than 10 M   (Smith et al.2003b) in a single eruptive event and survive, while the present-daystar and its binary companion present enduring challenges.The central mystery concerning  η  Car is the cause of its spec-tacular ‘Great Eruption’ in the mid-19th century (Davidson &Humphreys 1997) when it displayed erratic variability and brieflybecame the second brightest star in the sky despite its distance of  ∼ 2.3kpc (Smith 2006). Observing  η  Carinae at the Cape of GoodHope in the early to mid-19th century, J. F. W. Herschel first de-scribed the ‘sudden flashes and relapses’ of   η  Argus, as it wascalled at the time, and remarked that this star was ‘fitfully vari-able to an astonishing extent’ (Herschel 1847). At times it rivalledSirius and Canopus in brightness, but with an orange–red colour.Innes (1903) compiled a list of known 19th-century observations  E-mail: and published the familiar light curve that has been often repro-duced (e.g. Humphreys & Davidson 1994; Humphreys, Davidson& Smith 1999). The light curve was updated and corrected for scaleerrors by Frew (2004).The Homunculus nebula surrounding  η  Car is a prototypicalbipolar nebula, made famous in spectacular images from the  Hub-ble Space Telescope  (  HST  ; e.g. Morse et al. 1998). It had long beensuspected that the Homunculus srcinated from the Great Eruption(Gaviola 1950; Ringuelet 1958; Gehrz & Ney 1972), and proper-motion measurements of the expanding nebula later confirmed this,with estimated ejection dates of 1841 (Currie et al. 1996), 1844(Smith & Gehrz 1998) and 1846–48 (Morse et al. 2001). The ob-served historical brightening event and its expanding ejecta make  η Car uniquely valuable.Multiple eruptive episodes point to an enduring phase of in-stability, marked by repeated sequences of outburst and recovery.In addition to the multiple peaks during the Great Eruption thatwe discuss in this paper, the star brightened again around 1890whenitejectedanotherbipolarnebulacalledtheLittleHomunculus(Ishibashietal.2003;Smith2002,2005).Additionalnebulosityout-sidetheHomunculussuggestsmajorancienteruptions500–2000yr C   2011 The AuthorsMonthly Notices of the Royal Astronomical Society  C   2011 RAS  2010  N. Smith and D. J. Frew ago (Walborn, Blanco & Thackeray 1978; Walborn & Blanco 1988;Smith & Morse 2004; Smith, Morse & Bally 2005). The star hasalso been brightening in a non-steady way in modern times, with a jump in the 1940s (de Vaucouleurs & Eggen 1952) and again in thelate 1990s (Davidson et al. 1999).As our best studied example,  η  Car serves as the prototype for aclass of transient sources known variously as giant luminous bluevariable (LBV) eruptions, Type V supernovae (SNe), SN impostorsor  η  Car analogues, which are thought to represent non-terminaleruptions of massive stars. Smith et al. (2011) recently provided acomprehensive study and review of this class of objects. Although η  Car is often held as the prototype for this class, it is hardly atypicalcase.Itsmultiplepeaksandlongdurationareunusual(Smithet al. 2011). Unlike all extragalactic SN impostors, its nebula canbe studied in detail, and linking clues from the spatially resolvedejecta to the timing of brightening events is a critical piece of thepuzzle. Here we aim to provide the definitive historical light curveof   η  Car after reviewing all available historical documentation. Adetailed discussion of each observation in the historical light curveof   η  Carinae known at that time was given by Frew (2004). In thispaper, we collect and present 51 previously unpublished estimatesof the brightness of   η  Car from C. P. Smyth and Thomas Maclearin the 1840s. In combination with the data from Frew (2004), thisarchival data set of new measurements has allowed the first detailedlookatthephotometricbehaviourof  η Carinaeduringcriticalpointsin its Great Eruption of 1837–58. Including the new archival data,the character of the light curve is different from previous reportsconcerning the period of time centred around the peak of the GreatEruption, as we discuss below. 2 OBSERVATIONS2.1 New archival data The recent digital publication of the Royal Astronomical Soci-ety’s Herschel Archives has been a boon to historians of astronomy(Royal Astronomical Society 2004; see also Bennett 1977; Hoskin2005). An examination of this resource has revealed an importantnew data set of   η  Carinae’s brightness in 1842–43, compiled bySmyth (1843), which has now been fully reduced. This data setwas not summarized by Herschel (1847), and it was not availableat the time Frew (2004) compiled the historical light curve for  η Car. We also took the opportunity to examine relevant letters in theRoyal Society Archives that were available on microfilm from theUniversity Publications of America (Crowe 1990; Kesaris 1990).We perused letters from Thomas Maclear and C. P. Smyth (bothwere observers at the Cape of Good Hope) to Sir John Herschelbetween 1838 and 1865, in order to search for any additional un-published observations. The index of Crowe et al. (1998) permittedan efficient search through the archival letters (see Table 1).These new observations have cleared up some ambiguities andinconsistencies in the data summarized by Herschel (1847). Thediscrepancies pointed out by M¨uller & Hartwig (1918) have nowbeen clarified after examining these srcinal archival letters, ex-plained further in Section 2.2. We have reproduced a page from themanuscript of Smyth (1843) as Fig. 1 to illustrate the nature andscope of the source material. 2.2 Reduction procedure Following the procedure used by Frew (2004), all brightness esti-mates were reduced to the photopic visual system, with the zero-point equivalent to Johnson  V   for an A0 star. Almost all visual Table 1.  Summary of letters mentioning  η  Carinae from T. Maclear to J.F. W. Herschel between 1837 and 1865. Only letters making reference to η  Carinae are included. The column marked ‘CDK’ gives the referencenumber in the calendar of Crowe, Dyck & Kevin (1998). The last columnnotesifanextractoftheletterwaspreviouslypublishedinWarner&Warner(1984) or by Feast, Whitelock & Warner (1994).Reference Date Archive CDK OtherMaclear (1838a) 1838-1-24 RS:HS 12.183 3626 WW84Maclear (1838b) 1838-1-28 RS:HS 12.100 3630 WW84Maclear (1842) 1842-3-28 RS:HS 12.131 5140 –Maclear (1843) 1843-12-17 RAS:JH 3/1.1.2 5670 –Maclear (1844a) 1844-9-17 RAS:JH 3/1.1.13 5936 –Maclear (1844b) 1844-11-2 RS:HS 12.132 6005 –Maclear (1860) 1860-10-21 RS:HS 12.165 11449 FWW94Maclear (1863) 1863-5-17 RS:HS 12.168 12230 FWW94 observers naturally use photopic (foveal) vision, with an effectivewavelength λ eff   of 5600 Å, only slightly redward of Johnson  V   ( λ eff  of 5450 Å). We note that scotopic (peripheral or rod) vision ( λ eff    5100 Å) is considerably bluer than  V  , but is almost never used forstellarbrightnessestimation;seeSchaefer(1996b)andFrew(2004)for a fuller discussion.As before,  V   magnitudes for the comparison stars were takenfrom the Lausanne photometric data base (Mermilliod, Mermil-liod & Hauck 1997). No corrections for differential extinction wereapplied, as any factor is likely to be smaller than the adopted un-certainties of the visual magnitudes, derived from the interpolativemethod used.The brightness descriptions of Smyth (1843) are in the raw formand are equivalent to the traditional Argelander step method, wherethe difference in brightness of stars along a defined sequence isestimated.Smyth(1843)describedhisobservingmethodasfollows:‘The stars are here put down in their order of lustre as estimatedby the naked eye. The vertical strokes are intended to show thesupposed number of grades between any two.’An extract from his manuscript is reproduced here as Fig. 1, andwe use his data to determine the visual magnitude of   η  Carinaeby interpolation. We illustrate our reduction method using Smyth’sobservations for 1843 March 18; the values in parentheses are thegrades in brightness estimated by Smyth:Canopus (3)  η  Argus (2)  α  Centauri (3)  β  Centauri (1) α  Crucis (2)  β  Crucis (1)  γ   CrucisUtilizing the comparison star magnitudes given in table 3 of Frew (2004; see also Mermilliod et al. 1997), it can be seen that  η Carinae had  m V  = − 0.5  ±  0.2 on this date. We note that on thisnight,themagnitudeofeachgradeorstepwasnotconstantalongthesequence, ranging from 0.09mag between Canopus and α  Centaurito 0.37mag between  β  and  γ   Crucis. This is in fact typical of eachnight’s data. Using all of the data from Smyth’s manuscript leads usto adopt a mean step value of   m = 0.24 ± 0.13mag ( n = 160). Onsome nights (e.g. 1843 April 19)  η  Carinae was brighter than thebrightest comparison star, so the derived   m  value has been usedto determine the magnitude of   η  Carinae from extrapolation, witha larger uncertainty of 0.3mag adopted as a result (see also Frew2004).Ourderivedvisualmagnitudesaredenotedthroughoutby m V ,andrealisticuncertainties have been determined for each data point. Forthe majority of observers, there will be only a small colour term be-tweenthevisualsystemandJohnson V   formostnaked-eyestars,butthe difference between the  m V  and  V   systems for an emission-line C   2011 The Authors, MNRAS  415,  2009–2019Monthly Notices of the Royal Astronomical Society  C   2011 RAS   Light curve of Eta Car   2011 Figure 1.  Example presenting the srcinal source material, showing the list of comparisons by Smyth (1843). Reproduced from the Royal AstronomicalSociety Archives, MS Herschel (J) No 3/1.1.11. Reproduced by courtesy of the Royal Astronomical Society. star like  η  Carinae might be substantial. We do not currently haveenoughinformationtoquantifytheeffectsoftheemission-linespec-trum during the eruption, but we note that any error is likely to beless than the generous adopted uncertainty of  ± 0.2–0.5mag.TheobservationsofThomasMacleararereducedinasimilarway(Maclear 1842, 1843, 1844a,b). Maclear also used a step method,but his descriptions are more verbose and less precise. An examplefrom 1843 March 24 is typical (Maclear 1843):Decidedly not so brilliant as Canopus, brighter than  α 1 , 2 Centauri.From this description, the concluded magnitude is m V  =− 0 . 5 ± 0 . 2. Another example is Maclear’s observation of 1842 Mar 19.Maclear (1842) wrote, was considerably less than Rigel, less than  α  Crucis &much greater than  α  Hydrae. C   2011 The Authors, MNRAS  415,  2009–2019Monthly Notices of the Royal Astronomical Society  C   2011 RAS  2012  N. Smith and D. J. Frew The qualitative description and large difference in brightness be-tween Rigel ( β  Ori,  V   = 0.15) and  α  Hydrae ( V   = 1.98) precludean accurate estimate for  η  Car in this case. An approximate magni-tude of 1.0 ± 0.5 is inferred, since  η  Car was somewhat closer inbrightness to Rigel. The brightness of   α  Crucis sets an upper limitof   m V  =+ 0.75. Since a brightness ‘grade’ is ∼ 0.2mag, we againconclude that  η  Car had  m V  =  1.0 on this date. Importantly, thisobservation clarifies a discrepancy first noted by M¨uller & Hartwig(1918). Herschel (1847) had mistakenly recorded the wrong date(1843 Mar 19) for this observation when compiling his summaryof the available data. The summary table in Herschel (1847, p. 36)records the observation as being made on 1842 Mar 19, but thespecific wording in the text of Herschel (1847, p. 35) led M¨uller &Hartwig (1918) and Frew (2004) to assume that the year was in fact1843. Furthermore, the extensive series of observations recorded inthe letter by Maclear (1843) includes no such date. This error byHerschel led later workers to conclude that  η  Car underwent a fastdip and recovery in early 1843 (e.g. Li et al. 2009). The light curvepresented in Figs 2 and 3 corrects this error.Some of the other data have been modified from Frew (2004).In some cases, the descriptions of Herschel (1847) were found tobe a brief summary of the data contained in the srcinal letters.The new, srcinal estimates of Maclear and Smyth have alloweda greater density of points to be plotted during the crucial earlystages of the 1843 brightening (see Table 2). Another importantrevision in the light curve concerns the brightness during 1844 (seepanel b of Fig. 3). Herschel (1847) had only partially quoted fromMaclear’s description, leading Frew (2004) to incorrectly assumethat η  Car was marginally (‘scarcely’) fainter than Canopus in 1844September, with an estimated magnitude of   m V  =− 0 . 6. However,looking at the srcinal text in Maclear (1844a) shows clearly thathe estimated  η  Car as mid-way in brightness between  α 1 , 2 and  β Centauri (i.e.  m V  = + 0 . 2 ± 0 . 4). Maclear further stated that ‘thestar is and has been for 8 months stationary, & scarcely as bright asCanopus.’ We can hence infer that the star was apparently constantin brightness at  m V = 0.2 ± 0.4 from ∼ 1844.05 to 1844.71. 2.3 Results Table 2 summarizes the new  m V  magnitudes derived here. Thecolumns sequentially list the observer,  UT  date (as decimal year),the derived apparent visual magnitude, the adopted error on themagnitude and the reference from which the magnitude is derived.Iftheintervalbetweencomparisonstarmagnitudesislarge,thentheadopted uncertainty on the magnitude of   η  Car will be greater thanthenominaluncertaintyof  ± 0.2mag.Inthesecases,theuncertainty Figure 2.  The historical light curve for  η  Carinae. Panel (a) shows the full historical light curve from Frew (2004) in blue, with limits in grey arrows. Panel(b) zooms in on the Great Eruption during 1822–64. During this time interval, the previous light curve from Frew (2004) is in blue (points and dotted lines),while the revised light curve with new archival data that we discuss in this paper appears as black dots with error bars. Notes about the apparent colour arelisted above the light curve. The orange vertical dashes show predicted times of periastron passage if one simply extrapolates back from the currently observedorbital cycle with a stable 2022.7 d period (Damineli et al. 2008), whereas the red hash marks are similar but with a shorter (95 per cent) period before 1848.The dashed red horizontal line shows the quiescent magnitude of   η  Car as it would appear with zero bolometric correction. C   2011 The Authors, MNRAS  415,  2009–2019Monthly Notices of the Royal Astronomical Society  C   2011 RAS   Light curve of Eta Car   2013 Figure 3.  Same as Fig. 2, but zooming in on the events in (a) 1837–38, (b) 1843–45 and (c) the so-called ‘Lesser Eruption’ around 1890. is usually taken to be half the difference between the comparisonstar magnitudes. For ease of use, we have included all observationsbetween 1842 and 1845 in Table 2, and these data supersede themagnitudes presented by Frew (2004). For observations before andafterthisperiod,thereadershouldconsultthetabulateddatainFrew(2004).The first definitive observation of the variability of   η  Car camefromtheexplorerandnaturalistWilliamBurchellin1827(seeFrew2004 for a full account of Burchell’s observations). Writing fromBrazil on 1827 July 17, Burchell described it as ‘now of the firstmagnitude, or as large as  α  Crucis’ (Herschel 1847).The star was monitored between 1834 and 1838 by Sir JohnHerschel at the Cape of Good Hope (Herschel 1847). From 1834to 1837, the star was essentially constant, with  m V  =  1 . 2 ± 0 . 2(Frew 2004). Assuming a distance of 2300pc (Smith 2006), ( m −  M  ) 0 = 11.8mag, A V = 1.4mag,  M  bol =− 12.0magandabolometriccorrection of zero at the maximum light (consistent with an F-typephotosphere), we expect an apparent magnitude of  m V   1 . 2 out of eruption. The observed brightness before 1838.0 is very consistentwith this estimate (see Fig. 2).The Great Eruption is widely considered to have begun at theclose of 1838 when Herschel noted a rapid brightening of  ∼ 1magoveraperiodoflessthantwoweeks(Frew2004).Thestarthenfadedoverthefollowingmonths,butunfortunatelywehavenotrecoveredany observations between late 1838 and 1841, so there may havebeen other short-duration peaks in brightness that were missed, orthe star may have faded considerably (but see below). In 1842, themagnitude was approximately as it was prior to the commencementof the Great Eruption; our estimate is  m V  = 1.0 ± 0.4mag. It wasabout 0.5mag brighter in early 1843 when the brightness suddenlyincreased. The brightness peaked at about  m V  = − 0.8  ±  0.2magin late 1843 March. The star again faded in subsequent weeks, andfor most of 1844 it was constant at  m V  =  0.2  ±  0.2mag. At theclose of 1844, the star again brightened, and by 1845 January ithad reached  m V =− 1.0 ± 0.3mag, which is brighter than Canopus( V   =− 0.74).As described by Frew (2004), there is good evidence for markedfluctuations in brightness (amplitude up to 1mag on time-scales of days to weeks) during the Great Eruption. The brightening eventin 1843 March/April was remarkable (see observations of Maclearand Smyth described above), as was the brightening at the closeof 1844. After 1846, the observed variations were superposed on aslow decline (Frew 2004), with fluctuations noted by Jacob (1849)and Gilliss (1855, 1856). Between 1846 and 1856,  η  Car faded atan approximate rate of 0.1 magyr − 1 . It was still a star of the firstmagnitude at the close of 1857, before the rate of fading suddenlyincreasedby1859.Thismaybeduetotheonsetofdustcondensationfrom the stellar wind, or the Great Eruption may have ceased.Nearly all contemporary reports during the Great Eruption de-scribe  η  Car as ‘reddish’ or ‘ruddy’ (e.g. Mackay 1843; Smyth1845; Jacob 1847; Gilliss 1856; Moesta 1856; Abbott 1861;Tebbutt 1866), these observers sometimes make direct comparison C   2011 The Authors, MNRAS  415,  2009–2019Monthly Notices of the Royal Astronomical Society  C   2011 RAS
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks