Optical, UV, and EUV Oscillations of SS Cygni in Outburst

0.61 MB·English

by  Christopher W. Mauche

Optical, UV, and EUV Oscillations of SS Cygni in Outburst1,2
Christopher W. Mauche
LawrenceLivermoreNationalLaboratory,L-473,7000EastAvenue,Livermore,CA94550
4
0 Abstract. Iprovideareviewofobservationsintheoptical,UV(HST),andEUV(EUVEandChandraLETG)oftherapid
0 periodic oscillations of nonmagnetic, disk-accreting, high mass-accretion ratecataclysmic variables (CVs),with particular
2 emphasisonthedwarfnovaSSCyginoutburst.Inaddition,Idrawnattentiontoacorrelation,validovernearlysixorders
ofmagnitudeinfrequency, betweenthefrequenciesofthequasi-periodicoscillations(QPOs)ofwhitedwarf,neutronstar,
n and black hole binaries. This correlation identifies the high frequency quasi-coherent oscillations (so-called “dwarf nova
a oscillations”)ofCVswiththekilohertzQPOsoflowmassX-raybinaries(LMXBs),andthelowfrequencyandlowcoherence
J
QPOsofCVswiththehorizontalbranchoscillations(orthebroadnoisecomponentidentifiedassuch)ofLMXBs.Assuming

3 that the same mechanisms produce the QPOs of white dwarf, neutron star, and black hole binaries, this correlation has
2 importantimplicationsforQPOmodels.

1
v
INTRODUCTION CVs are in many ways better laboratories in which to
4
8 studyaccretionprocesses(themass-accretionratevaries
4 Rapid periodic oscillations have been studied in cata- systematicallybythreeordersofmagnitudeduringdwarf
1 clysmic variables (CVs; [29]) since the early 1970s in novaoutbursts,thepeakluminositiesarewellbelowthe
0 the optical and the early 1980s in the extreme ultravi- Eddingtonvalue, radiationpressure is unimportant,and
4
olet (EUV) and soft X-rays [25, 20, 30]. (Nominally) relativisticeffectsareinsignificant),becauseCVoscilla-
0
nonmagnetic, disk-accreting, high mass-accretion rate tionscanbestudiedfromthegroundintheopticalwith
/
h (“high-M˙;” novalike variables and dwarf novae in out- modest-size telescopes, and because CVs offer unique
p burst) CVs manifest “dwarf nova oscillations” (DNOs) diagnostics of the oscillations in the optical, UV, and
o- withperiodsP≈3–30sandhighcoherence(Q≈104– EUV. Since the date of this meeting, Warner [30] has
r 106), as well as “quasi-periodic oscillations” (QPOs) suppliedan excellentreview of the oscillationsof CVs,
t
s with longer periods (by a factor of ≈ 13, see §3) and particularly those in the optical. This contribution con-
a far lower coherence (Q ≈ 1–10). In contrast, the os- centrates on observations of oscillations in the optical,
:
v cillations of intermediate polars/DQ Her stars [21], the UV, and EUV flux of the best-studied CV, the dwarf
i white dwarfanaloguesof accretion-poweredX-raypul- nova SS Cyg. During outburst, oscillations have been
X
sars,havelongerperiods(P=33–7190s)andfarhigher detected in its optical [22, 5, 4, 20], UV (HST) [§4],
ar coherence(Q≈1010–1012).Whilethepulsationperiods and EUV/soft X-ray (HEAO 1, EXOSAT, ROSAT, and
ofintermediatepolarsvarygradually(|P˙|<10−10ss−1), EUVE)[3,2,6,10,28,16,14]fluxwithperiodsranging
those of dwarf novae vary significantly during dwarf from3sto11s.
nova outbursts, decreasing on the rising branch and in-
creasingonthedecliningbranchoftheoutburst.
Although QPO research in compact binaries largely SIMULTANEOUS OPTICAL, UV,AND
shifted to LMXBs in the mid-1980s, the study of rapid
EUV OBSERVATIONS
oscillations in CVs continuesto be of value because of
themanysimilaritiesbetweenCVsandLMXBs,because
Mauche & Robinson [16] and Wheatley, Mauche,
& Mattei [34] have described simultaneous optical
(AAVSO visual magnitudes and McDonald Observa-
1 DedicatedtomycolleagueandfriendJanetAkyuzMattei,whothis tory 2.7 m telescope high-speed UBVR photometry),
yearcelebratesher30thyearasdirectoroftheAmericanAssociation EUV (EUVE DS and SW: l ≈70–120 Å), and X-ray
ofVariableStarObservers.
2 Based in part on observations with the NASA/ESA Hubble Space (RXTE PCA: E ≈ 2–15 keV) observations of a nar-
TelescopeobtainedattheSpaceTelescopeScienceInstitute,whichis row asymmetric outburst of SS Cyg in 1996 October.
operatedbytheAssociationofUniversitiesforResearchinAstronomy, The resulting optical, EUV, and X-ray light curves
Incorporated,underNASAcontractNA5-26555.
are shown in the upper panel of Figure 1. They can
Optical,UV,andEUVOscillationsofSSCygniinOutburst February2,2008 1
FIGURE 2. Period of the EUV oscillation as a function of
DS count rate during the 1996 October outburst of SS Cyg.
Points on the rising branch of the outburst are distinguished
with crosses. Grey lines are the unweighted fits to the data:
P=7.26I−0.097 sand P=2.99I−0.021 s. Pointsenclosed by
thedottedboxareplottedattwicetheobservedperiods.
orderthevirialtemperaturekT =GM m /3R ∼10
vir wd p wd
keV),andthenintheEUVwhentheboundarylayerbe-
comes optically thick to its own radiation and becomes
relatively cool (of order the blackbody temperature
kT =k[GM M˙/8ps R3 ]1/4∼10eV).Thedelaybe-
FIGURE1. Upperpanel:AAVSOoptical,EUVE,andRXTE bb wd wd
tweentheriseoftheopticalandEUV/X-rayemissionis
light curves of the 1996 October outburst of SS Cyg. EUVE
adirectmeasureofthetimeittakesfortheheatingwave
DS and SW measurements are shown respectively by the
filledcirclesandsquareswitherrorbars,RXTEPCAmeasure- tosweepthroughthedisk.Thedelaysofapproximately
ments(kindlysuppliedbyP.Wheatley)areshownbytheopen 11,11,and 3 daysbetweentheriseoftheopticalandthe
2 4 4
squares, individual AAVSO measurements are shown by the EUVfluxofoutburstsofSSCyg,UGem,andVWHyi,
smalldots,andthehalf-daymeanopticallightcurveisshown respectively,areconsistentwithaheatingwavevelocity
by the histogram. Intervals of observations at McDonald Ob- r /t ≈3kms−1 [15].
servatoryareindicatedbythethickbars.Lowerpanel:Oscilla- disk delay
Asshownin thelowerpanelof Figure1, oscillations
tionperiodversustime.EUVEDSandMcDonaldObservatory
opticalmeasurementsareshownbytheopencirclesandfilled were detected in the EUVE DS count rate light curves
starred diamonds, respectively. Points enclosed by the dotted duringanintervalofapproximatelyoneweekduringthis
boxareplottedattwicetheobservedperiods. outburst.Theoscillationwasfirstconvincinglydetected
ontherisingbranchoftheoutburstataperiodof7.81s,
itfellto6.59soveranintervalof4.92hr(Q=1.5×104),
be understood in the context of the thermal-viscous jumpedto2.91s,andthenfellto2.85soveraninterval
instabilitymodelofdwarfnovaoutbursts[8,26].When of4.92hr(Q=3.0×105) beforeobservationswith the
thedisksurfacedensityreachessomecriticalvaluefirst DSwereterminated.WhenDSobservationsresumed3.4
in the outer disk, the disk plasma is heated locally and dayslaterduringthedecliningbranchoftheoutburst,the
a heating wave is launched though the disk, heating periodoftheoscillationwasobservedtorisefrom6.73s
the disk plasma and increasing its viscosity. As this to 8.23 s overan intervalof 2.10 days(Q=1.2×105).
material sinks toward the white dwarf, it converts its When the EUV oscillation period was approximately
gravitational potential energy into rotational kinetic 2.9sanditsamplitudewas25–30%,aconservativeupper
energy and radiation. This radiation comes out first in limitfortheX-rayoscillationamplitudewas7%[34].
the optical and then in the UV as hotter parts of the ItisclearfromFigure1thattheperiodoftheEUVos-
disk are activated (in steady state, Tdisk (cid:181) r−3/4). When cillationofSSCyganticorrelateswiththeDScountrate,
this material reaches the boundary layer between the beinglongwhenthecountrateislowandshortwhenthe
disk and the surface of the white dwarf, it converts its countrateishigh.Toquantifythistrend,Figure2shows
prodigiousrotationalkinetic energyinto radiation.This the log of the period of the oscillation as a function of
radiationcomesoutfirstinX-raysbecausetheboundary the log of the DS count rate. As in the previous figure,
layer is initially optically thin and hence quite hot (of the data fall into two groups: one during the early rise
Optical,UV,andEUVOscillationsofSSCygniinOutburst February2,2008 2
FIGURE3. ChandraLETGspectrumofSSCygobtainedon2001January16.LabelsidentifytheH-andHe-likelinesofC,N,
O,Ne,Mg,andSiandthestrongestlinesinthemodeloftheEUVspectrum.Inthelowerpanel,thedataareshownbytheblack
histogram,theabsorbedblackbodycontinuumbythesmoothgraycurve,theindividualionspectrabythegraycurves,andthenet
modelspectrumbythethickgraycurve.Thestrongestlinesinthemodelarelabeled.
(distinguishedwithcrosses)anddeclineoftheoutburst, became “stiffer” by a factor of approximately 5. This
the other during the interval after the frequency of the evolution is consistent with SS Cyg pulsating at a fun-
oscillationhaddoubled.Thetrendsduringtheearlyrise damentalperiodP >∼ 6.5s,thenswitchingtoafirsthar-
and decline of the outburst are the same and can be fit monic and stiffening its period-intensity (by inference,
byafunctionoftheformP=P0I−a (whereI istheDS period-M˙) relationship so as to avoid oscillating faster
count rate) with P0 =7.26 s and a =0.097, consistent than 2Pmin ≈5.6 s. This minimum period is consistent
with the trendsobservedduring outburstsof SS Cyg in withtheKeplerianperiodattheinneredgeoftheaccre-
1993Augustand1994June/July[10]. tion disk if the mass of the white dwarfM ≈1.1M
wd ⊙
Despite any direct evidence that the white dwarf in (and the Nauenberg [18] white dwarf mass-radius rela-
SSCyg ismagnetic,itisusefulto considertherequire- tionapplies).A securevalueofthewhitedwarfmassis
mentsofamagnetosphericmodeltoexplaintheperiod- neededtoconfirmthisinterpretation.
intensity(byinference,period-M˙)relationshipobserved As shown in the upper panel of Figure 1, the Mc-
inSSCyg.ForastarwithadipolemagneticfieldB(r)= DonaldObservatoryUBVRphotometrywasobtainedon
m /r3 (where m = B(R⋆)R3⋆ is the dipole moment), the three consecutive nights during the early decline of the
disk is truncated at a radius r (cid:181) m 4/7M˙−2/7, hence the outburst. Oscillations in the optical flux were detected
0
Keplerian frequency n (r )= 1 (GM /r3)1/2 (cid:181) M˙3/7. on the second and third nights with periods of 6.58 s
K 0 2p ⋆ 0
With a =3/7=0.286, this is far “softer” than the re- and6.94s,respectively,consistentwiththecorrespond-
lationship observed in SS Cyg. For a multipole mag- ingEUV oscillationperiods.Duringthese intervals,the
netic field B(r) = m/rl+2 (where m = B(R )Rl+2 is amplitude of the EUV oscillation was 34%, while in
l l ⋆ ⋆
themultipolemoment),r (cid:181) m4/(4l+3)M˙−2/(4l+3),hence UBVR the amplitudes were 0.11%, 0.07%, 0.05%, and
0 l
n (r ) (cid:181) M˙3/(4l+3). Under these assumptions, SS Cyg 0.07%, respectively. On the third night, during two in-
K 0
tervals of strictly simultaneous optical and EUV data,
requiresa surface magneticfield strengthB(R )∼0.1–
⋆
the periods and phases of the oscillations were deter-
1MG(whichissufficientlylowtobehard-to-impossible
mined to be the same within the errors. The phase dif-
todetectdirectly),anda high-ordermultipolefield(l=
7+4)[10]. ferenceD f 0=0.014±0.038impliesD t=0.10±0.26s
−2 for P=6.94 s. The 3s upper limit D t ≤0.88 s corre-
As restrictive as this is, the situation is worse during
spondstoadistancer=cD t≤2.6×1010cm.Assuming
the peak of the outburst: after the oscillation frequency
thattheEUVoscillationoriginatesnearthewhitedwarf
doubled,theperiod-intensityrelationshipcanbefitwith
P=2.99 s and a =0.021: not only did the oscillation and that the optical oscillation is formed by reprocess-
ingofEUVfluxinthesurfaceoftheaccretiondisk,the
frequencydouble,it’s dependenceon the DScountrate
Optical,UV,andEUVOscillationsofSSCygniinOutburst February2,2008 3
delayD t=r(1−sinicosj )/c,wherethebinaryinclina-
tioni≈40◦ and0≤j ≤p istheazimuthalanglefrom
the line of sight. Then, the distance to the reprocessing
site r=cD t/(1−sinicosj )≤1.6×1010 cm, whichis
about 30 white dwarf radii or one-third the size of the
disk.
CHANDRALETG OBSERVATIONS
Mauche [13, 14] presented photometric and spectro-
scopic results from a Chandra Low Energy Transmis-
sion Grating (LETG) observation of SS Cyg obtained
on 2001 January 16 during the plateau phase of a wide
asymmetric outburst. The resulting spectrum, shown in
Figure3,containsemissionlinesofH-andHe-likeC,N,
O,Ne,Mg,andSiandFeL-shellionsintheX-rayband,
and a quasi-continuum extending from 40 Å to 130 Å
in the EUV band. The EUV spectrum can be modeled
as an absorbedblackbodywith flux scattered outof the
line of sight by scores of ground-state transitions of a
broadrangeofionsinthesystem’soutflowingwind.The
model fit shown in the figure has the following param- FIGURE 4. Power spectra of Chandra LETG l =42–120
eters: blackbodytemperaturekT =21.5 eV, neutralhy- Å count rate light curves of SS Cyg in outburst. (a) Power
drogencolumndensityN =5.0×1019cm−2,fractional spectrum of 47 ks of data binned to 1 s time resolution. (b)
H
emitting area f = 4.5×10−3 (hence luminosity L = Meanpowerspectrumof47consecutive1kslightcurves.(c)
Meanpowerspectrumofthe471kslightcurvesafterscaling
4p fR2wds T4=2×1033ergs−1),andwindvelocityV = bythevaryingfrequencyofthen 1≈0.11Hzoscillation.Note
2500kms−1andmass-lossrateM˙ =3×10−11M⊙yr−1. the simultaneous presence of oscillations at n 0≈0.0090 Hz,
Thestrongestlinesinthemodelarelabeledinthefigure n 1 ≈0.11 Hz (n 0/n 1 ≈0.088), and possibly a third at n 2 ≈
andincludeO V–VI,Ne V–VIII,Mg V–X,Si VI–IX, n 0+n 1≈0.12Hz.
S VII–VIII, and intermediate change states of L-shell
Fe.
Power spectra were calculated from background- extendington 2≈n 0+n 1≈0.12Hz.Thisfeatureisre-
subtracted count rate light curves constructed from the vealed as a distinct peak in the power spectrum of the
± first-order LETG event data. The X-ray and EUV harder(l =42–70Å)halfoftheEUVcomponentofthe
components of the spectrum were isolated by applying LETG spectrum.In EUVE DS data obtainedduringthe
l =1–42 Å and l =42–120 Å wavelength filters, re- 1994June/JulyoutburstofSSCyg,oscillationswerede-
spectively.Consistentwiththe1996OctoberEUVEand tected at n 0 ≈0.012 Hz and n 1 ≈0.13 Hz (and its first
RXTE results, the optically thick EUV component of harmonic2n 1),indicatingthepresenceofoscillationsat
the LETG spectrum oscillates, while the optically thin periods P0 ≈83 s and P1 ≈7.7 s [12]. The ratio P1/P0
X-ray componentdoes not. The power spectrum of the for the ChandraLETG and EUVE DS powerspectra is
EUV lightcurve,showninthe upperpanelofFigure4, 0.088and0.096,respectively.
manifests excess power at frequencies n ≈0.0090 Hz SSCygisnottheonlycompactbinaryinwhichmul-
0
and n ≈ 0.11 Hz, indicating the presence of oscilla- tiple periodicitieshave been detected. Woudt& Warner
1
tionsatperiodsP ≈110sandP ≈9.1s.Toinvestigate [35] discuss a number of instances when multiple peri-
0 1
theseoscillationsmoreclosely,thesoftX-raylightcurve odicitiesweredetectedintheopticalfluxofVWHyiin
wasdividedinto47consecutive1ksintervals.Although outburst. During the decline of the 2000 February out-
the powerspectra of these light curvesare rather noisy, burst, DNOs with periods PDNO = 27–37 s and QPOs
the n 1 ≈0.11 Hz oscillation typically appears as a sin- with periods PQPO =400–580 s were detected simulta-
gle D n =0.001 Hz peak with a frequency in the range neously. The ratio PDNO/PQPO = 0.064–0.071 is simi-
n =0.109–0.112Hz.Themeanofthese47powerspec- larto theperiodratiosobservedinSSCyg.Inaddition,
1
traisshowninthemiddlepanelofFigure4.Inaddition Psaltis, Belloni,&vanderKlis[24] showedthatinfive
tothepeaksatn ≈0.0090Hzandn ≈0.11Hz,there Z sources a tight correlation exists between the “hori-
0 1
appears to be a shoulder on the higher frequency peak zontal branch oscillation” (HBO) frequency n HBO and
Optical,UV,andEUVOscillationsofSSCygniinOutburst February2,2008 4
FIGURE 5. n –n correlation for neutron star binaries
high low
(opencircles),blackholebinaries(filledcircles),andthewhite
dwarf binaries SS Cyg (filled diamonds) and VW Hyi (open
squares).TheSSCygdataarefromMauche[14],theVWHyi
dataarefromWoudt&Warner[35],andtheneutronstarand
blackholebinarydataarefromBelloni,Psaltis,&vanderKlis
[1],andwerekindlysuppliedbyT.Belloni.Dottedlinedrawn
throughthepointsisn low=0.08n high. FcuIrGveUsRoEf 6th.e U19p9p9erJupnaeneol:utEbUurVstEoafndSSAACVygS.ODoSptmicaelasluigreh-t
mentsareshownbythefilledcircleswitherrorbars;individual
AAVSOmeasurements areshownbythesmall dots;half-day
the frequencyn of the lower frequencymember of the meanoptical light curve isshown bythehistogram. Intervals
l
pair of “kilohertz” (kHz) QPOs, with n /n ≈0.12. of HST observations are indicated by the thick bars. Lower
HBO l
Furthermore, by identifying with n and n the fre- panel:Oscillationperiodversustime.EUVEDSandHST UV
HBO l measurementsareshownbytheopencirclesandstarredfilled
quenciesofvarioustypesofpeakednoisecomponentsin
diamonds,respectively.
atollsources,otherneutronstarbinaries,andblackhole
binaries, they (andsubsequentlyBelloni, Psaltis, & van
der Klis [1]) extendedthis correlationovernearly three Hz for a M⋆ = 1.4M⊙ neutron star with r >∼ 3RS =
orders of magnitude in frequency. Figure 5 shows this
6GM /c2 =12.4 km, as required by general relativity.
correlationandshowsthattheEUVdataofSSCygand ⋆
In addition to their frequencies, the DNOs of CVs and
theopticaldataofVWHyiextendthiscorrelationnearly
thekHzQPOsofneutronstarbinariesaresimilarinthat
two orders of magnitude lower in frequency. This con-
theyhaverelativelyhighcoherenceandhighamplitudes,
nectionbetweentheoscillationsofCVsandLMXBshas
their frequency scales with the inferred mass-accretion
since beenstrengthenedbyWarner,Woudt,& Pretorius
rate, and they sometimes occur in pairs. Assuming that
[33],whousedpublished,archival,andnewdatatoadd
the same mechanisms produce the oscillations in white
a number of other CVs to the correlation, extending it
dwarf,neutronstar,andblackholebinaries,thedataex-
another order of magnitude lower in frequency (to the
cludetherelativisticprecessionmodelandthemagneto-
magnetic disk-accreting dwarf nova GK Per, for which
n is the QPO frequencyand n is the white dwarf spheric and sonic-pointbeat-frequencymodels (as well
low high
asanymodelrequiringthepresenceorabsenceofastel-
spinfrequency).
larsurfaceormagneticfield)[14].
This connectionbetween the oscillations of CVs and
LMXBsidentifiestheDNOsofCVswiththekHzQPOs
ofLMXBs,andtheQPOsofCVswiththeHBOs(orthe
HST STISOBSERVATIONS
broad noise component identified as such) of LMXBs.
Note that the frequencies of the DNOs of CVs and the
kHzQPOsofneutronstarbinariesaresimilarinthatthey The final campaign to be discussed is one in which
are comparable to the Keplerian frequency at the inner EUVEandHST SpaceTelescopeImagingSpectrograph
edgeoftheaccretiondiskof,respectively,awhitedwarf (STIS)observationswereobtainedduringanoutburstof
and neutronstar: n ≤0.14 Hz for a M =1M white SS Cyg in 1999 June. The resulting optical and EUV
K ⋆ ⊙
dwarf with r ≥ R⋆ = 5.5×108 cm, while n K <∼ 1570 lights curves are shown in the upper panel of Figure 6.
Unfortunately, the outburst proved to be both short, so
Optical,UV,andEUVOscillationsofSSCygniinOutburst February2,2008 5
FIGURE7. Upperpanel:HST STISE140HspectrumofSSCygobtainedon1999June10.86UT.LabelsidentifytheCIVP
Cygniprofile,theHe II emissionline,andvariousstronginterstellarabsorptionlines.Unitsoffluxdensityare10−11 ergcm−2
s−1Å−1.Lowerpanel:EventsfoldedontheoscillationperiodP=9.28sforechelleorders(a)1–6(l ≈1495–1528Å),(b)8–11
(l ≈1533–1557Å:CIV),(c)13–23(l ≈1562–1629Å),and(d)25–26(l ≈1634–1648Å:HeII).
onlytwooftheHSTvisitscoincidedwiththepeakofthe werecalculatedforlightcurvesofeachechelleorder,and
outburst, and weak, so the peak DS count rate was rel- excess power was detected in the neighborhood of the
atively low. EUV oscillations were detected in the DS CIVandHeIIlines,butnotinthecontinuum.Todeter-
count rate light curves during an interval of approxi- minethewavelengthdependenceoftheoscillationwhile
mately21 daysnearthepeakoftheoutburst.Thelower maximizingthesignal,powerspectrawerecalculatedfor
2
panel of Figure 6 shows the evolution of the period of lightcurvesproducedfromeventsfromechelleorders1–
the EUV oscillation, which gradually fell from ≈10 s 6(l ≈1495–1528Å),8–11(l ≈1533–1557Å:CIV),
to ≈8.75 s, and then rose back to ≈10 s. Four orbits 13–23(l ≈1562–1629Å),and25–26(l ≈1634–1648
of HST observationswere obtained duringeach of four Å: He II). The lower panels of Figure 7 show the rel-
visitson1999June8,10,14,and18.Duringeachvisit, ative light curves for each of these channels folded on
STIS was used in TIME-TAG mode with (1) the NUV- the9.28soscillationperiod.Byfittingasinusoidalfunc-
MAMA with the G230L grating (l ≈ 1600–2800 Å, tion A+Bsin2p (f −f ) to the foldedlightcurves,the
0
one orbit), (2) the FUV-MAMA with the G140L grat- relative oscillation amplitudes B/A were found to be
ing (l ≈ 1150–1600 Å, one orbit), and (3) the FUV- 0.13%±0.06%, 0.60%±0.08%, 0.04%±0.06%, and
MAMAwiththeE140Hechellegratingcenteredatl = 0.53%±0.15%, respectively: the oscillation was con-
1598 Å (l ≈1495–1690Å, two orbits). Unfortunately, vincinglydetectedintheCIV andHe IIchannels,only
itwasnecessaryforreasonsofdetectorsafetytousethe weakly(at2.2s )detectedinthel ≈1495–1528Åcon-
F25NDQaperture/neutraldensityfilterwiththefirsttwo tinuumchannel,andnotdetectedinthel ≈1562–1629
configurations.UVoscillationsweredetectedduringtwo Å continuum channel. This result is in contrast to the
consecutiveorbitson the early decline from outburstin eclipsing dwarf nova OY Car in superoutburst, which
lightcurvesproducedfromtheechellegratingeventdata showedoscillationsintheUVcontinuum,butnotinthe
(2×107 photons per orbit!) with periods of 9.06 s and C IV emission line [9]. The result for SS Cyg doesnot
9.28 s, consistent with the corresponding EUV oscilla- exclude the possibility that the oscillation amplitude of
tionperiods. theUVcontinuumwasashighasisobservedintheop-
More interesting is the spectrum of the UV oscilla- tical(0.05%–0.1%,§2).Thepossibilityremainsthatthe
tions. The upper panel of Figure 7 shows the spectrum opticaloscillation,liketheUVoscillation,isdominated
derivedfromtheechellegratingobservationobtainedon by the lines. In support of this, note that in the dwarf
1996 June 10.86 UT. It consists of a bright continuum novaV2051Ophinoutburst,themeanoscillationampli-
onwhichare superposedtheC IV l 1549PCygnipro- tudeoftheBalmeremissionlinesisapproximatelytwice
file, the He II l 1640 emission line, and various inter- thatoftheopticalcontinuum[27].
stellar absorption lines (the discontinuities in the spec- How are we to understand these results? Numerous
trumevery≈6Åareduetosmallerrorsintheassumed linesofevidenceidentifythewhitedwarf/boundarylayer
effective areas of the 32 echelle orders). Power spectra astheprimarysourceoftheoscillationsobservedinnon-
Optical,UV,andEUVOscillationsofSSCygniinOutburst February2,2008 6
magnetic,disk-accreting,high-M˙ CVs.Thepulsationsof areavailableforstudyintheoptical,UV,andEUV.The
intermediate polars and the 360◦ phase shifts of eclips- growthareainthisfieldofresearchinthenearfuturewill
ing intermediate polars [31], novalike variables [17, 7], likely be high-speedspectroscopicobservationsof CVs
anddwarfnovaeinoutburst[20]areinterpretedinterms withlargeopticaltelescopes,suchasKeckandtheVLT.
ofamodelinwhichrotatingbeamsofEUV/X-raypho-
tonsare reprocessedinthe surfaceoftheaccretiondisk
[19, 23]. Warner & Woudt [32] argue that DNOs are ACKNOWLEDGMENTS
causedbyan equatorialbelton thesurfaceof the white
dwarf that is first spun up and then spun down during IthankmycollaboratorsJ.Mattei,E.Robinson,J.Ray-
dwarfnovaoutbursts,andthatQPOsarecausedbyver- mond,andP.Wheatleyfortheircontributionstothisre-
ticalthickeningsoftheinnerdisk,whichalternatelyob- search, the AAVSO for target-of-opportunityalerts and
scureandreflectradiationfromthecentralsource. optical data, P. Wheatley for supplying the RXTE data
TheHeIIl 1640Balmera linearisesfromthesurface shown in Fig. 1, and T. Belloni for supplying the neu-
of the accretion disk and is a probe of EUV flux, as tron star and black hole binary data shown in Fig. 5.
it arises from the cascade following photoionization of Support for this work was provided in part by NASA
HeIIbyEUV(l <228Å)photons.TheCIVfeature,on through(1)ChandraAwardNumberGO1-2023Aissued
theotherhand,isformedbyscatteringofl =1548,1550 bytheChandraX-rayObservatoryCenter,whichisop-
ÅphotonsinSSCyg’soutflowingwind.Itissomething erated by SAO for and on behalf of NASA under con-
ofasurprisethatthislinepulsatesatall,giventhelarge tract NAS8-39073and (2) grantnumberGO-06545.01-
volumeinwhichitisformed.Evenmoresurprisingisthe 95AfromSTScI,whichisoperatedbytheAURA,under
apparentlackofaphaseshiftacrosstheCIVprofile:the NASAcontractNAS5-26555.Thisworkwasperformed
blueandredsidesoftheprofileriseandfallinphase.To undertheauspicesoftheU.S.DepartmentofEnergyby
investigatetheimplicationsofthisresult,weconstructed University of California Lawrence Livermore National
a simple modelin which photonsfrom a centralsource LaboratoryundercontractNo.W-7405-Eng-48.
were scattered into the line of sight by a spherically
symmetric constant-velocity wind, taking into account
thelighttraveldelaysforthescatteredphotons.Different REFERENCES
pulse-phaseresolvedprofileswereobtainedforasource
with a rotating one-armed beam, a rotating two-armed 1. Belloni,T.,Psaltis,D.,&vanderKlis,M.2002,ApJ,572,
beam,andanisotropicsourcewhosefluxissinusoidally 392.
modulated. The synthetic profiles showed the greatest 2. Córdova,F.A.,Chester,T.J.,Mason,K.O.,Kahn,S.M.,
phaseshiftbetweentheblueandredsidesoftheprofile &Garmire,G.P.1984,ApJ,278,739.
3. Córdova,F.A.,Chester,T.J.,Tuohy,I.R.,&Garmire,
for the rotating one-armed beam source and the least
G.P.1980,ApJ,235,163.
phase shift for the modulated isotropic source, but a
4. Hildebrand,R.H.,Spillar,E.J.,&Stiening,R.F.1981,
moredetailedinvestigationisrequiredtodetermineifthe ApJ,243,223.
rotatingtwo-armedbeamsourcecanbeexcludedbythe 5. Horne,K.,&Gomer,R.1980,ApJ,237,845.
data. 6. Jones,M.H.,&Watson,M.G.1992,MNRAS,257,633.
7. Knigge,C.,Drake,N.,Long,K.S.,Wade,R.A.,Horne,
K.,&Baptista,R.1998,ApJ,499,429.
8. Lasota,J.-P.2001,NewAstronomyReviews,45,449.
CONCLUSION
9. Marsh,T.R.,&Horne,K.1998,MNRAS,299,921.
10. Mauche,C.W.1996,ApJ,463,L87.
AsIhopethiscontributionhasdemonstrated,thefuture 11. Mauche,C.W.1997,ApJ,476,L85.
of the study of the rapid periodic oscillations of CVs 12. Mauche,C.W.1997,inAccretionPhenomenaandRelated
Outflows,ed.D.T.Wickramasinghe,L.Ferrario,&G.V.
is bright. Additionalobservationsof additionalsystems
Bicknell(SanFrancisco:ASP),251.
can help establish the proposed equivalence of CV and
13. Mauche,C.W.2002,BAPS,47,#N17.118.
LMXB QPOs, and will provideuniqueandquantitative 14. Mauche,C.W.2002,ApJ,580,423.
tests of QPO models. Data can be obtained from the 15. Mauche, C.W.,Mattei,J.A.,&Bateson,F.M.2001,
groundintheopticalandfromspaceintheUVandEUV, inEvolutionofBinaryandMultipleStarSystems,ed.
the system parameters (binary inclination, white dwarf Ph.Podsiadlowski,S.Rappaport,A.R.King,F.D’Antona,
&L.Burderi(SanFrancisco:ASP),367.
mass, radius, and rotation velocity) can be measured,
16. Mauche,C.W.,&Robinson,E.L.2001,ApJ,562,508.
eclipse mapping in edge-on systems allows the sites of
17. Nather,R.E.,&Robinson,E.L.1974,ApJ,190,637.
fluxmodulationstobelocatedanddissected,dwarfnova
18. Nauenberg,M.1972,MNRAS,254,493.
outburstsprovideadramaticandsystematicvariationin 19. Patterson,J.1979,ApJ,234,978.
the mass-accretion rate, and diagnostic emission lines 20. Patterson,J.1981,ApJS,45,517.
21. Patterson,J.1994,PASP,106,209.
Optical,UV,andEUVOscillationsofSSCygniinOutburst February2,2008 7
22. Patterson,J.,Robinson,E.L.,&Kiplinger,A.L.1978,
ApJ,226,L137.
23. Petterson,J.A.1980,ApJ,241,247.
24. Psaltis,D.,Belloni,T.,&vanderKlis,M.1999,ApJ,520,
262.
25. Robinson,E.L.1976,ARA&A,14,119.
26. Schreiber,M.R.,Hameury,J.-M.,&Lasota,J.-P.2003,
A&A,410,239.
27. Steeghs,D.,O’Brien,K.,Horne,K.,Gomer,R.,&Oke,
J.B.2001,MNRAS,323,484.
28. vanTeeseling,A.1997,A&A,324,L73.
29. Warner,B.1995,CataclysmicVariableStars(Cambridge:
CUP).
30. Warner,B.2004,PASP,inpress(astro-ph/0312182).
31. Warner,B.,Peters,W.L.,Hubbard,W.B.,&Nather,R.E.
1972,MNRAS,159,321.
32. Warner,B.,&Woudt,P.A.2002,MNRAS,335,84.
33. Warner, B., Woudt, P. A., & Pretorius, M. L. 2003,
MNRAS,344,1193.
34. Wheatley,P.J.,Mauche, C.W.,&Mattei,J.A.2003,
MNRAS,345,49.
35. Woudt,P.A.,&Warner,B.2002,MNRAS,333,411.
Optical,UV,andEUVOscillationsofSSCygniinOutburst February2,2008 8