Temporal resolution of a pre-maximum halt in a Classical Nova: V5589 Sgr observed with STEREO HI-1B

0.21 MB·

by  S. P. S. Eyres

MNRAS000,1–6(2017) Preprint16March2017 CompiledusingMNRASLATEXstylefilev3.0
Temporal resolution of a pre-maximum halt in a Classical
STEREO
Nova: V5589 Sgr observed with HI-1B
S. P. S. Eyres,1⋆ D. Bewsher,1 Y. Hillman,2 D. L. Holdsworth,1 M. T. Rushton,3,1
D. Bresnahan,1 A. Evans,4 P. Mr´oz5
7 1Jeremiah Horrocks Institute, Universityof Central Lancashire, Preston PR1 2HE, UK
1 2Department of Geosciences, Raymond and Beverly SacklerFaculty of Exact Sciences,Tel-AvivUniversity,Tel-Aviv69978, Israel
0 3Astronomical Institute of the Romanian Academy, Str. Cutitulde Argint 5, 040557, Bucuresti, Romania
2 4Astrophysics Group, Keele University,Keele, Staffordshire ST5 5BG, UK
5Warsaw UniversityObservatory,Al. Ujazdowskie 4, 00-478 Warszawa, Poland
r
a
M
Accepted2017January31.Received2017January26;inoriginalform2016December22

5
1
ABSTRACT

] Classicalnovaeshowa rapidriseinopticalbrightnessovera few hours.Untilrecently
R
the rise phase, particularly the phenomenon of a pre-maximum halt, was observed
S sporadically. Solar observation satellites observing Coronal Mass Ejections enable us
h. to observethe pre-maximumphasein unprecedentedtemporalresolution.We present
p observationsofV5589SgrwithSTEREO HI-1Batacadenceof40min,thehighestto
- date.Wetemporallyresolveapre-maximumhaltforthefirsttime,withtwoexamples
o
each rising over 40 min then declining within 80 min. Comparisonwith a grid of out-
r
t burst models suggests this double peak, and the overall rise timescale, are consistent
as with a white dwarf mass, central temperature and accretion rate close to 1.0 M⊙,
[ 5×107 K and10−10 M⊙ yr−1 respectively.The modelling formallypredicts massloss
onsetatJD2456038.2391±0.0139,12hrsbeforeopticalmaximum.Themodelassumes

3 amain–sequencedonor.Observationalevidenceis forasubgiantcompanion;meaning
v
the accretion rate is under–estimated. Post–maximum we see erratic variations com-
6
monlyassociatedwithmuchslowernovae.Estimatingthedeclineratedifficult,butwe
2
0 place the time to decline two magnitudes as 2.1 < t2(days) < 3.9 making V5589 Sgr
a“very fast”nova. The brightest point defines“day 0”as JD 2456038.8224±0.0139,
9
0 althoughatthishighcadencethemeaningoftheobservedmaximumbecomesdifficult
. to define. We suggest that such erratic variability normally goes undetected in faster
1
novaeduetothelowcadenceoftypicalobservations;implyingerraticbehaviourisnot
0
necessarily related to the rate of decline.
7
1 Key words: novae, cataclysmic variables – stars:individual:V5589Sgr
:
v
i
X
r
a 1 INTRODUCTION to the absolute magnitude via the maximum-magnitude-
rate-of-decline relationship (MMRD; Downes & Duerbeck
Classical novae are interacting binary stars, in which a
2000), for both Galactic and extra–galactic populations
white dwarf (WD) accretes matter from a (usually) main–
(Darnley et al. 2006). Slower novae often show erratic vari-
sequence–like companion, generally in a close orbit with a
ations and remain bright for longer times than is consis-
period of a few hours to days. This results in a build up of
tent for the MMRD, when considered in the light of in-
material on the WD surface, leading to thermonuclear run-
dependent distance measurements. The relationship is also
aways (TNRs) that cause an outburst, at ultraviolet (UV),
not useful for novae that decline most rapidly, as the ab-
optical and infrared (IR) wavelengths in a few hours, and
solute magnitude becomes insensitive to decline time (e.g.
theejection of around10−5 to10−4M⊙ of matterat veloci- Capaccioli et al. 1990).
ties of a few 100 to a few 1000 km s−1 (Warner 2008). De-
cline then follows over the course of weeks or months, with
Recently observations and theoretical developments
the time to drop 2 or 3 magnitudes from maximum related
haveestablished that,priortoopticalmaximumX–ray and
even gamma ray emission is seen (Ackermannet al. 2014).
Due to the rate of ascent, the pre–maximum optical light–
⋆ E-mail:[email protected](SPSE) curvehasbeenonlyrarelyobserved,andatcadencesof1−2
(cid:13)c 2017TheAuthors
2 S. P. S. Eyres et al.
datapointsperday.Someshowapre–maximumhalt(PMH) R2/R⊙ ∼ 0.11Phr ∼ 4.2R⊙, confirming that the secondary
that had not been subjected to theoretical modelling. is evolved. The last pre-eruption image was taken on 2012
Hounsell et al.(2010)presentedfourlight–curvestaken April 20.3728 UT at I = 16.90. These authors also note
with theSolar Mass Ejection Imager (SMEI)down totime thatwiththislongerperiod,thecompanionislikelytobea
resolutionsof80min,showingthePMHtobeubiquitousin subgiant.Weston et al.(2016)cometothesameconclusion.
theirsample.FurtheranalysisofSMEIlight-curvesofnovae
waspresentedbyHounsellet al.(2016),withacatalogueof
14 systems. They derived peak time, maximum magnitude
3 OBSERVATIONS
andanestimateoftherateofdeclineforallsystems,andten-
tativelydetectedPMHsintwooftheirsample.Hillman et al. The Solar TErrestrial RElations Observatory (STEREO)
(2014)havedemonstratedthattheformofthesehaltsvaries consistsoftwonearlyidenticalspacecraft,inheliocentricor-
withthegrossparametersofWDmass,WDcentraltemper- bits,oneaheadoftheEarth(STEREO-A)andtheotherbe-
ature, and theaccretion rate from thecompanion. hindtheEarth(STEREO-B)(Kaiser et al.2008).Thetwin
HerewepresentdatafromtheSTEREO HI-1Bcamera spacecraftwerelaunchedon2006October26,andthesepa-
for V5589 Sgr, following the outburst from the initial rise, ration betweeneach spacecraft andtheEarth hasincreased
through the optical peak to the point where it fades be- by approximately 22◦ per year since achieving helocentric
lowdetectabilityforthisinstrument.Thecadenceisaround orbits.
40min,makingthesethehighesttime-resolutionobservation TheSunEarthConnectionCoronalandHeliosphericIn-
of this phase to date. We discuss the rise phase, including vestigation (secchi) instrument suite (Howard et al. 2008)
the PMHs, in the context of the models of Hillman et al. on STEREO includes the heliospheric imagers (HI-1
(2014).Wealsocontrastthepost–maximumbehaviourwith and HI-2). The HI instruments have been used to ob-
muchslowerClassicalNovae(CNe)andsuggestthattheem- serve a variety of variable stars including classical nova
pirical split in this behaviour between slow and fast novae V5583 Sgr (Holdsworth et al. 2014), and long period vari-
is in part a function of the typical one day sampling of the ables (Wraight et al. 2012).
light–curves. Data from the HI-1instument on STEREO-B (HI-1B)
are used in this paper. HI-1B has a field-of-view (FOV) of
20◦, which extendsfrom ∼4−24◦ elongation, with a pixel
size of ∼ 70 arcsec, (Eyles et al 2009). The STEREO HI-
2 V5589 SGR
1Binstrument’smainpassbandcoverstherangefrom600−
V5589 Sgr(NovaSgr2012, PNV J17452791–2305213, coor- 750nm,butalsohascontributionsat300−450nm andfrom
dinates J2000 17 45 28.03 −23 05 22.7) was discovered on 900−1100nm (see Fig. 6 of Bewsher et al. 2010)
2012 April 21.0112 UT (Sokolovsky et al 2012) at V ∼9.6, The HI-1B images consist of 30 exposures of 40s each,
rising to a reported maximum of V = 8.8 (Korotkiy et al. withasummedimage cadenceof40min(Eyleset al2009).
2012)by2012April21.654UT(CBAT2015).Pre–discovery Each exposure is scrubbed of cosmic ray hits prior to sum-
images detected the outburst on 2012 April 20.8403 UT ming (Howard et al. 2008). The summed images constitute
down to V=10.2 (CBAT 2015). theL0.5fitsfileswhichareavailabletodownloadfromsev-
Early spectroscopy indicated high velocity ejecta in eral websites including the UK Solar System Data Centre
this nova, with line widths (FWHM) of 5600 km s−1 (UKSSDC1).
(Korotkiy et al. 2012) to 6500 km s−1 (Esipov et al 2012). TheL0.5datahavebeenpreparedwiththesswidlpro-
Observations in X-ray (Sokolovsky et al 2012; Nelson et al cedure, secchi prep. This routine applies the shutterless
2012a) and radio (Nelson et al 2012a) wavelengths shortly correctiontothedata,thelargescaleflatfielddeterminedby
afteropticalpeakshowednon–detections,andsuggested no Bewsher et al. (2010), and the pointing and optical param-
shocked component in the expanding ejecta at this early eter calibration of Brown, Bewsher & Eyles (2009). It also
stage of thedecline. identifies missing blocks and saturated pixels, filling these
Swift UVOT detected the object on 2012 April 25.7 at pixelswith NaN values.
aUVM2brightnessof13.71±0.09mag,fadingrapidlyfrom Asdiscussedin Bewsher et al. (2010),thepointspread
13.8 to 13.5 over a 5975s observation (Nelson et al 2012b). function (PSF) of the HI-1B instrument is well approxi-
By 2012 May 10 hard X–ray emission was present, consis- mated with a Gaussian profile, and the aperture size of 3.1
tent with a column of N(H) = (3±1)×1021 cm−2 and a pixelssuggested byBewsher et al. (2010) is used to extract
plasmatemperatureofatleast2.7×108K;theUVM2bright- thenovalight–curvefromthedatausingaperturephotome-
ness was 16.0±0.2 mag by this date (Nelson et al 2012b). try.ThesolarF-coronaisthedominantsourceintheHI-1B
Weston et al. (2016) argue that the radio and X-ray char- data, and is removed usinga daily runningbackground.
acteristics are consistent with high–speed, low–mass ejecta Toextractthelight–curvepresentedinthispaper,each
that accelerates over thefirst few weeks after outburst. HI image is processed with secchi prep, and the position
ThenovahasbeenobservedbytheOGLEsurveysince ofthenovadeterminedusingthepointinginformationgiven
2010 (Mro´z et al. 2015). In the quiescent light–curve they in the header of the fits file. Aperture photometry is per-
found eclipses with a period of 1.59230(5) days, which they formed on this position, and the number of counts (DN/s)
interpretastheorbitalperiod.Sucharelativelylongorbital recorded. This process is repeated for every HI image from
period suggests that thesecondaryis a subgiant.Ifthesec- 2012 April 18 to 2012 May 7, where the nova crosses the
ondaryfillsitsRochelobetheorbitalperiodcanbeusedto
estimate the radius of the secondary (see e.g. King 1989).
UsingEq(2.3.10) ofKingwefindthatthesecondaryradius 1 http://www.ukssdc.ac.uk/
MNRAS000,1–6(2017)
Resolved pre-maximum halt in the nova V5589 Sgr 3
CCD of the instrument. The counts recorded are then con- the decline time from peak is used to determine the speed
verted intoa magnitudeusing thefollowing equation classofCNe(Payne-Gaposhkin1957;Duerbeck2008),which
inturncorrelateswiththeenergeticsoftheevent,itisuseful
I
M =−2.5log tot (1) to identify this point. There is one brightest point in the
HI µF
(cid:18) (cid:19) STEREO data, albeit less than 1 σ above adjacent points.
whereM istheHImagnitude,I istherecordedcounts, ThispointisatJD2456038.8224±0.0139, andcanbetaken
HI tot
µisacalibrationconversionfactorandF isthefluxofVega as thepeak magnitude.
(Bewsher et al.2010),thezero–pointofthemagnitudescale. Determiningthetimetodeclinebytwomagnitudes,t2,
Theerroron µF,σ ,isestimated tobeoftheorderof 5%, andhencethespeedclass,issomewhatmoreuncertain.The
F
therefore, brightness at peak is 7.94±0.09, and the brightness drops
σ below 9.94 for the first time after JD 2456040.9057±0.0139
F =0.05 (2) butrecovers 2 h later. Thelast timeit is brighterthan this
µF
is the point at JD 2456042.3224. The average magnitude
ItisassumedthattheuncertaintiesonI ,σ arePois-
tot I binned over 18 points (12 h interval) reaches this level in
son counting statistics, and the conversion from DN (units
thebincentredonJD2456041.7668. Thesegiveat2 timeof
of I ) to photoelectrons is 1 DN = 15 photoelectrons
tot between2.1and3.9days,placingthenovafirmlyinthe“very
(Eyles et al 2009),therefore, fast”class. The STEREO passband incorporated more red
I lightthantheBorVbandswheret2 isusuallydetermined.
σI = 1t5ot. (3) Nonethelessthevaluesarewellwithinthe“veryfast”range,
r and there is an expectation that this object will be in the
Theuncertaintyonthemagnitudemeasurementisthen upperrangeforintrinsicluminosityandvelocityoftheejecta
given by ofthenovaevents.Thisisconsistent withtheidentification
of a high–speed shock byWeston et al. (2016).
−2.5 σ 2 σ 2
σ = I + F (4)
MHI (cid:12)(cid:12)ln(10)s(cid:18)Itot(cid:19) (cid:18)µF(cid:19) (cid:12)(cid:12)
(cid:12) (cid:12)
(cid:12)(cid:12) 2 (cid:12)(cid:12) 4.1 Pre–maximum halts
(cid:12) −2.5 I1to5t (cid:12)2
= (cid:12) v +(0.05) (cid:12) (5) A key finding of our work is the presence in the rise–
(cid:12)ln(10)uqI  (cid:12)
(cid:12) u tot (cid:12) phase of deviations from a monotonic increase that are
(cid:12) u (cid:12)
(cid:12) t  (cid:12) consistent with PMHs seen in other classical novae. For
(cid:12) (cid:12)
Thisvaluewasc(cid:12)alculatedonapoint–by–pointb(cid:12)asisand V5589Sgrwehaveidentifiedtwoshortdeviations–peaking
(cid:12) (cid:12)
isincludedinFig.1aserrorbars.Theerrorduetotheback- atJD2456037.8502±0.0139 andJD2456037.9891±0.0139 –
ground subtraction has not been included in the errors on withthisphenomenon.FromFig.1wecanseethattherise
the light–curve. However, it is estimated that theerror due to peak occurs between one point and the next (i.e. within
to thebackground subtraction is of theorder of 10% of the 40min),withthedeclineoverthreepoints(i.e.over80min),
calculated error for a handful of points, and less than 5% both with the same shape within the uncertainties. As the
for the majority of points. This is due to the background PMHs appear V5589 Sgr moves between pixels 80 and 110
subtraction being calculated using the average over a days along the 540th row of the CCD, so the nova is well away
worthofobservationsleadingtocomparativelysmalluncer- from theedge of the1024×1024 pixel detector.
tainties. Theinitialrisereflectsachangeofmorethan3σpoint-
In thefollowing we use the40 min cadence of thedata to-pointasaPMHstarts(uncertaintiesdeclinewithincreas-
as the basis for uncertainties in any dates or intervals in ingbrightness).ThePMHdeclineistoalevelwithin2σ of
time. This translates to ±0.0139 d, with the third decimal the pre–rise magnitude. Modelling by Hillman et al. (2014)
place being conservative given that observations are made suggests that PMHs are often short–lived excursions above
at thiscadence towithin a few seconds. the general upwards trend of the rise. We explore this by
taking linear interpolations between the “before” and “af-
ter”points for both features. These are found to follow a
lineconsistentwiththesubsequentrise.Thusthetwopeaks
4 RESULTS
are consistent with being such excursions from the general
The light–curve of V5589 Sgr is presented in Fig. 1. We upwardstrend,ratherthanbeingdefinedbydipsthatinter-
see the rise to maximum, and then a rather erratic decline. rupttherise.Thecombination ofthechangerelativetothe
We also overlay American Association of Variable Star Ob- uncertaintiesandthedeviationfromwhatonewouldexpect
servers(AAVSO)B,Vandvisualestimatedataforcompari- from backwards extrapolation of the rise, allows us to be
son.UncertaintiesinAAVSOvisualestimatesareestimated confident that they can be reasonably described as PMHs
as 0.1 mag although this varies with location and experi- as modelled byHillman et al. (2014).
ence of the contributing observers. We can see that while
the STEREO and AAVSO data are consistent, the latter
has a much lower cadence and relies primarily on visual es-
timates.
5 DISCUSSION
Thetimeofmaximumisakeycharacteristicdetermined
for previous novae, and often used to refer to subsequent We discuss the pre-maximum halts during the rise phase,
development,withthebrightestpointlabeledas“day0”.As and thenthe behaviourfollowing theoptical peak.
MNRAS000,1–6(2017)
4 S. P. S. Eyres et al.
Figure1.STEREOHI-1Bdatawitherrorbars,plusAAVSOvisualestimates(triangles),V(squares)andB(invertedtriangles)included
for comparison; offset by 1 mag upwards for clarity. STEREO is approximately R-band, for which data are available in AAVSO only
afterJD2456051 (MJD 56051.5), about 6dafter theendofthisdatarun.VerticaldashedlinesmarkthePMHsdescribedinthetext.
Theinsetzoomsinonthe∼14hourperiodaroundthePMHs.
5.1 Double pre–maximum halt the point when convection ceases to be efficient near the
surface causing a reduction in the energy flux. The opacity
Duringtherisetomaximum(Fig.1),weseetwoverysimilar then decreases allowing radiation to become the dominant
features that represent departures from the smooth, mono- energy transfer mechanism and reverse the dip. This starts
tonic rise normally considered to be the usual behaviour todrivemass loss.
for CNe before the optical peak. We have interpreted these
The model agrees less well with the behaviour of
asthesometimes–observedPMHs.Whentakenincombina- V5589 Sgr once convection ceases to dominate evolution
tion with those seen by Hounsellet al. (2010) we can con-
around the PMHs. Nonetheless we have constrained the
cludethattheseeventsarecommon,andhavebeengenerally
onset of mass loss to occur between the peak of the first
missed in historical observations. PMH and the maximum of the optical light (i.e. between
In Fig. 2 we compare a model from the grid of JD 2456037.8502 and JD 2456038.8224), with the model
Hillman et al.(2014)withtheSTEREO data.Wehavecho- formally giving this as JD 2456038.2391. The modelling
sen parameters which give the best qualitative agreement assumes a main–sequence donor star. While the subgiant
withthedata.Wehavenotattemptedtofittheparameters donor would have a lower H abundance, this simply means
tothedata,butinsteadweofferthisassupportthatthegen- the accretion rate derived from the model is an underesti-
eraldevelopmentofthePMHsduringtheriseisunderstood. mate of the true rate. An optical spectrum in quiescence
Themodel plot shows therelevanttime period from Fig. 4, would be valuableto confirm thenatureof the secondary.
Panel 4, plot 100.50.10 of Hillman et al. (2014). The model
starts from well before the span of the data, and to assess
thesimilarity,wehavesimplyalignedthepeakofthesecond
5.2 Post–maximum erratic decline
halt.Byinspectionwecanthenseethattheduration,mag-
nitudeand interval between the halts is consistent between There are a number of excursions visible in Fig. 1 signif-
the model and the data. The PMHs occur in the model at cantly above or below the mean magnitude during the de-
MNRAS000,1–6(2017)
Resolved pre-maximum halt in the nova V5589 Sgr 5
6 (2003) discuss V723 Cas as an example, and suggest these
6.5 mstaarstss loss “flare-like”features might be due to interactions within the
ejecta or mass transfer bursts from the companion, as pro-
7
posed by Chochol, O’Brien & Bode (2000). By contrast er-
7.5
raticbehaviourisnotgenerallyseeninfastexamplesofCNe,
8
where the rise and fall is seen to be smooth. However the
mag8.5 vastmajorityofground–basedlight–curvesaremadeupofa
9 numberofobservationseverynight,withconsiderablevaria-
tionintheconsistencyofobservationsandaremainlyvisual
9.5
estimates. The data presented here provide unprecedented
10 STEREO
stability and point-to-point reliability at a regular 40 min
Model
10.5 cadence.
11 Hounsell et al. (2010) found some erratic behaviour
-1.1 -0.95 -0.8 -0.65 -0.5 -0.35 -0.2
early on in V1280 Scoand toa lesser extentin KT Eri, the
days
second also being very fast. As noted in their Section 4 the
Figure2.Comparisonofdata(lowerline)andmodel(upperline) declinetimeandtheformofthePMHsdonotseemtohave
fortheparametersgivingtheclosestmatchtothepre–maximum anycorrelation;V5589Sgrshowsadifferentformagaineven
behaviour(WDmass1.0M⊙,centraltemperature5×107 Kand thoughtheinitialdeclineratewassimilartothatofKTEri.
massaccretionrate10−10 M⊙yr−1).Thetimeaxisisrelativeto We suggest that the association of erratic brightness varia-
day0atJD2456038.8224.
tions near peak with slower novae is a feature of the poor
cadenceofmost light–curvedatatodate.Itisclearly afea-
cline after day 0. We consider if any of this structure is ture of some examples of fast novae, but at this point we
consistentwithperiodicity.Wealsoexaminehowsignificant do not have enough examples to judge if it is less common
theexcursions are given theoverall scatter in thedata. than in slower novae. Hounsell et al. (2016) showed further
An orbital period of 1.59230 d determined by examplesofnon-smoothdeclinesfrommaximuminmanyof
Mr´oz et al. (2015) is based on the ephemeris of HJD = their novae light curves, however due to the limiting mag-
2455000.724±0.015+(1.59230±0.00005)×E fortheeclipses. nitudeand contamination thepresence of erratic variations
To test for periodicities in the decline phase, we ex- may not be astrophysical.
cluded data obtained before day 0 and perform a discrete In fact, oncewe start to record theearly light–curveof
Fourier transform on the remaining data using the Pe- novae at the cadence achieved here, the definition of“day
riod04 software of Lenz& Breger (2005). After fitting and 0”needs to be revisited. An alternative approach to deter-
removing the slope of the decline, we search for significant mining the peak in Fig. 1 is to treat all fluctations from
periodsinthede-trendeddata.Giventhequalityoftheob- a smooth rise and decline as anomalous. In this case some
servations,werelaxthe‘standard’S/Nlimitof4(e.g.Koen form of fit to the overall rise and decline might be used to
2010) to 3 to determine if a peak in the FT is significant, give an estimate of the peak that is derived in a manner
however we find no peaks in the FT above this limit. Fur- consistent withthosefor novaewith onlylow cadencedata.
thermore, there is no evidence in our data of the orbital Asecond–orderpolynomialfitthroughthepeakhasamaxi-
period determined byMr´oz et al. (2015). mumof8.16±0.10mag,at56038.68,or160minearlierthan
This is consistent with post–outburst Optical Gravi- the absolute brightest measurement used to define“day 0”
tational Lensing Experiment (OGLE) observations show- in Section 4. Thus at this high cadence, where we see fluc-
ing that eclipses do not start occuring again until HJD = tationsontimescalesofafewhours,therelevanceofoptical
2456062.82544 orpossibly aslateasHJD=2456076.92252. maximum and hence“day 0”starts to become problematic.
Prior to this the entire binary must be within the photo- Howeveras STEREO and otherfacilities allow usto exam-
sphere of the nova ejecta, and so must be larger than the ineanincreasingnumberoffastnovaeinthisdetail,wemay
orbitalradiusof∼7.2R⊙×(M/2M⊙)1/3 forasystemmass beable to reassess thespeed class concept.
of M.
While there are a number of variations in the decline
phase, none are consistently separated by this interval, or
6 CONCLUSIONS
a multiple thereof. In addition a simple polynomial fit to
the decline suggests only the minima centred around MJD Wehavecollectedanopticallight–curveofthefastClassical
56039.7668, 56040.1002 and 56040.9335 are consistent with NovaV5589Sgratatime–resolution of40min,thebestfor
anything other than the overall level of fluctuations seen. anynovatodate.Thisdatashowtherisetomaximumwith
None of these are separated by the period in Mr´oz et al. pre–maximum halts, and the subsequent somewhat erratic
(2015). Thus whatever the origin of these dips, neither decline,in greater detailthan seen before. Wefindthat the
eclipses nororbital modulation appear toexplain them. halts are consistent with a binary model having a WD of
The speed class appears to be related both to the lu- mass 1.0 M⊙, central temperature 5×107 K and mass ac-
minosity of the nova at the peak, and the kinetic energy of cretionrate10−10 M⊙ yr−1.Asthedonorstarisasubgiant,
the ejecta, with both being higher for faster novae. Much thelastoftheseislikelyanunderestimate.Themodelshows
lessconsistently,moreslowlydecliningexamplesoftenshow thePMHsarepre–cursorstotheonsetofmasslossandfor-
erratic brightening and dimming after the initial rise, pre- mally predicts this as starting at JD 2456038.2391±0.0139
sentingmultiplefalsemaximaandmakingthedetermination for V5589 Sgr.
ofthepeak,andhencethedeclinerate,difficult.Evanset al. We suggest that the erratic decline behaviour, a com-
MNRAS000,1–6(2017)
6 S. P. S. Eyres et al.
mon characteristic of slow novae, is missed in previous ob- Kaiser.M.L.,etal.,2008,SpaceScienceReviews,136,5
servations of some faster examples by the typical daily ob- KingA.R.,1989,in:ClassicalNovae,BodeM.F.,EvansA.(eds),
servations. Our work, and that of Hounsell et al. (2010), pp16–33, J.Wiley&Sons,Chichester,UK
demonstrate that some fast novae also demonstrate erratic KoenC.2010,Ap&SS,329,267
Korotkiy S., et al. 2012, Central Bureau Electronic Telegrams,
behaviour close to optical peak. This is not understood or
3089,1
well modelled at this time. We also find that at such high
LenzP.,&BregerM.2005,CommunicationsinAsteroseismology,
cadence, the common identification of the peak as“day 0”
146,53
becomes increasingly difficult and possibly misleading.
Mr´ozP.,UdalskiA.,PoleskiR.,etal.,2015,ApJS,219,26
STEREO–A continues to operate and we will retrieve NelsonT.,MukaiK.,SokoloskiJ.,ChomiukL.,RupenM.,Mio-
data of further CNe that are serendipitously bright as they duszewski,A.,2012a,ATel.4088
cross one or more of the four camera apertures. This will NelsonT.,MukaiK.,SokoloskiJ.,ChomiukL.,RupenM.,Mio-
allow usto investigate theprevalence and natureof PMHs, duszewski,A.,2012b,ATel.4110
and any erratic behaviourclose tooptical maximum. Payne-Gaposhki, C., 1957, The Galactic Novae. Amsterdam:
North-Holland.
Sokolovsky K., Korotkiy S., Elenin L., Molotov I., 2012, ATel.
4061
ACKNOWLEDGEMENTS WarnerB.,2008,in:ClassicalNovae,BodeM.F.,EvansA.(eds),
pp16–33, CambridgeUniversityPress,Cambridge
Data created during this research is openly WestonJ.H.S.,etal.,2016,MNRAS,460,2687
available from the University of Central WraightK.T.,etal.,2012,MNRAS,426,816
Lancashire data repository, UCLanData, at
http://doi.org/10.17030/uclan.data.00000049. Thispaper has been typeset fromaTEX/LATEX fileprepared by
The Heliospheric Imager (HI) instrument was devel- theauthor.
oped by a collaboration that included the Rutherford
Appleton Laboratory and the University of Birmingham,
both in the United Kingdom, and the Centre Spatial de
Li´ege (CSL), Belgium, and the US Naval Research Labora-
tory (NRL),Washington DC, USA. The STEREO/secchi
project is an international consortium of the Naval Re-
search Laboratory (USA), Lockheed Martin Solar and As-
trophysics Lab (USA), NASA Goddard Space Flight Cen-
ter (USA),Rutherford Appleton Laboratory (UK),Univer-
sity of Birmingham (UK),Max-Planck-Institutfu¨rSonnen-
systemforschung (Germany), Centre Spatial de Li`ege (Bel-
gium),Institutd’OptiqueTh´eoriqueetAppliqu´ee(France),
and Institutd’AstrophysiqueSpatiale (France).
DLHacknowledgesfinancialsupportfromtheSTFCvia
grant ST/M000877/1.
REFERENCES
AckermannM.,etal.,2014,Sci,345,554
BewsherD.,etal.,2010,SolarPhysics,264,433
BrownD.S.,BewsherD.,EylesC.J.,2009,SolarPhysics,254,185
CBAT, 2015, ”Transient Object Followup Reports”
PNV J17452791-2305213, accessed 2015/10/27
http://www.cbat.eps.harvard.edu/unconf/followups/J17452791-2305213.html
Capaccioli M., della Valle M., D’Onofrio M., Rosino L., 1990,
ApJ,360,63
ChocholD.,O’BrienT.J.,BodeM.F.,2000,NewAR,44,P61
DarnleyM.J.,BodeM.F.,KerinsE.,etal.,2006,MNRAS,369,
257
DownesR.A.,DuerbeckH.W.,2000,AJ,120,2007
DuerbeckH.W.,2008,in:ClassicalNovae,BodeM.F.,EvansA.
(eds),pp1–15, CambridgeUniversityPress,Cambridge
EsipovV.F.,SokolovskyK.V.,KorotkiyS.,2012,ATel.4094
EvansA.,GehrzR.D.,GeballeT.R.,etal.,2003,AJ,126,1981
Eyles,C.J.,etal.,2009,SolarPhysics,254,387
Hillman Y., Prialnik D., Kovetz A., Shara M. M., Neill J. D.,
2014,MNRAS,437,1962
HoldsworthD.L.,etal.,2014,MNRAS,438,3483
HounsellR.,etal.,2010,ApJ,724,480
HounsellR.,etal.,2016,ApJ,820,104
HowardR.A.,etal.,2008,SpaceScienceReviews,136,67
MNRAS000,1–6(2017)