NASA Technical Reports Server (NTRS) 20130013626: Swift J045106.8-694803: A Highly Magnetised Neutron Star in the Large Magellanic Cloud

0.47 MB·English

by  NASA Technical Reports Server (NTRS)

Mon.Not.R.Astron.Soc.000,1–12(2012) Printed10December 2012 (MNLATEXstylefilev2.2)
Swift J045106.8-694803; a highly magnetised neutron star
in the Large Magellanic Cloud.
H. Klus1∗, E.S. Bartlett1, A.J. Bird1, M. Coe1, R.H.D. Corbet2 and A. Udalski3
2
1The Faculty of Physical and Applied Sciences, University of Southampton, Highfield, Southampton SO17 1BJ, United Kingdom
1
2Universityof Maryland Baltimore County, X-ray Astrophysics Laboratory, Mail Code 662, NASA Goddard Space Flight Center,
0
Greenbelt, MD 20771, USA
2
3Warsaw UniversityObservatory, Aleje Ujazdowskie 4, 00-478 Warsaw, Poland

c
e
D
Accepted 2012October 28.Received2012October 19;inoriginalform2012September 10

7

] ABSTRACT
E
H WereporttheanalysisofahighlymagnetisedneutronstarintheLargeMagellanic
. Cloud (LMC). The high mass X-ray binary pulsar Swift J045106.8-694803 has been
h
observedwithSwiftX-raytelescope(XRT)in2008,TheRossiX-rayTimingExplorer
p
(RXTE) in 2011 and the X-ray Multi-Mirror Mission - Newton (XMM-Newton) in
-
o 2012. The change in spin period over these four years indicates a spin-up rate of
r −5.01±0.06s/yr,amongstthe highestobservedfor anaccretingpulsar.This spin-up
t
s rate can be accounted for using Ghosh and Lamb’s (1979) accretion theory assuming
a it has a magnetic field of (1.2±0.2)×1014 Gauss. This is over the quantum critical
[ 0.7
fieldvalue.Thereareveryfewaccretingpulsarswithsuchhighsurfacemagneticfields

3 and this is the first of which to be discovered in the LMC. The large spin-up rate is
v consistentwithSwiftBurstAlertTelescope(BAT)observationswhichshowthatSwift
0 J045106.8-694803has had a consistently high X-ray luminosity for at least five years.
8 Opticalspectra have been used to classify the optical counterpartof Swift J045106.8-
6 694803 as a B0-1 III-V star and a possible orbital period of 21.631±0.005 days has
7
been found from MACHO optical photometry.
.
0
Key words: X-rays: binaries – stars: neutron - stars: binaries - stars: magnetar -
1
2 individual: Swift J045106.8-694803
1
:
v
i
X 1 INTRODUCTION pulsation period is equated with the rotation period of the
r neutronstar’scrustandthisvaluechangeswithtimeasac-
a HighmassX-raybinaries(HMXB)arebinarysystemscom-
creted matter transfers angular momentum to or from the
posedofeitheraneutronstar,whitedwarforblackholeand
star.
an opticalcompanion, eitherasupergiant star–inthecase
ofsupergiantX-raybinaries(SGXB)–oradwarf,subgiant HMXB are composed from massive, and therefore rel-
orgiantOBestar-inthecaseofBe/X-raybinaries(BeXB). atively short lived, stars and so their presence indicates
HMXB can be detected by the X-ray, optical and infrared that a new population of stars formed relatively recently
emission they produce. (Grimm, Gilfanov and Sunyaev 2003). This may be why
X-rays are produced as matter is transferred from the more HMXB have been discovered in the Milky Way and
optical star to its denser companion. It is accelerated as the Small Magellanic Cloud (SMC) than the Large Magel-
it moves into the gravitational potential well produced by lanic Cloud (LMC).
the denser star and - if the denser star is a neutron star The HMXB in the Milky Way are mostly found
- it is then channelled by the star’s magnetic field lines to within the spiral arms (Coleiro and Chaty 2011) and the
the magnetic poles. The accreted matter then rapidly de- SMC underwent a burst of star formation about 200 mil-
celerates when it reaches the surface and potential energy lion years ago when it was about three times closer to
is converted to heat which energises the plasma, producing the LMC than it is today (Gardiner and Noguchi 1996;
X-ray hot spots. These can appear as ‘beams’ if they peri- McBride et al 2010). The SMC has an HMXB popula-
odically travelpastourlineofsightasthestarrotates.The tion which is comparable in number to the Galaxy de-
spite being almost two hundred times less massive. Whilst
about half of the HMXB in the Galaxy are BeXB, all
∗ E-mail:[email protected] (HK) but one HMXB in the SMC is a BeXB system. The
(cid:3)c 2012RAS
2 H. Klus, E.S. Bartlett, A.J. Bird, M. Coe, R.H.D. Corbet and A. Udalski
ObservationID StartDate StartTime Exposure(ks) ObservationID StartDate StartTime Exposure(ks)
00038029001 2008-10-23 03:56:38 6.56 96441-01-01-00 2011-10-28 13:54:55.6 9.005
00038029002 2008-11-11 14:35:46 6.40 96441-01-01-01 2011-10-28 10:51:38.6 0.849
00038029003 2008-11-14 00:18:15 2.59
Table 2.SummaryofRXTEdatasets.
00038029004 2009-01-10 13:23:59 4.32
00038029005 2011-09-30 16:50:35 4.47
00038029006 2011-10-03 18:28:46 5.90
well as the intrinsic H I column density (NH) and pho-
Table 1.SummaryofXRTdatasets. ton index, were calculated using ftool xspec. The spec-
tra were described by an absorbed power law with a
fixed Galactic foreground column density of 8.4×1020 cm-2
LMC, on the other hand, contains relatively few HMXB,
(Dickeyand Lockman 1990) and abundances set in accor-
most of which are BeXBs (Negueruela and Coe 2002;
dance with Wilms, Allen and McCray (2000). Intrinsic ab-
Liu, van Paradijs and van den Hauvel 2005; Sturmet al
sorption and the abundances of elements heavier than he-
2012).
lium were set to 0.4 (Borkowski, Hendrickand Reynolds
Swift J045106.8-694803 is part of a newly discovered 2006).X-rayspectrawerethencompiledinxspec over0.2-10
BeXB system located in the LMC. It was first discov-
keV.
ered by Beardmore et al (2009) at an RA, Dec (J2000)
The 0.2-10 keV flux of each dataset was determined
of 04:51:06.8 and -69:48:03.2 respectively, with an uncer- usingxspec.Theluminositywasthencalculatedusingadis-
tainty of 3.5(cid:2)(cid:2). Beardmore et al measured an X-ray flux of tance of 50.6±1.6 kpc(Bonanos et al 2011).
(1.68±0.11)×10-11 erg cm-2s-1 over 0.3-10 keV, fit with a Thelightcurveofeachdatasetwasextractedinxselect
powerlaw ofphoton index0.96±00..0046 with acolumn density and a Lomb-Scargle normalised periodogram was produced
of (1.9±0.3)×1021 cm-2. for each using time-series analysis package Period3.
The 14-195 keV flux was found to be (2.8±0.3)×10-11
erg cm-2s-1 fit with a power law of photon index 2.5±0.4.
Beardmore et al also report a pulsation period of 187 sec- 2.1.2 Swift/BAT
onds.TheopticalcompaniontothisX-raysourceistheV=
14.70 bluestar[M2002] 9775 (Massey 2002) withanorbital The Swift Burst Alert Telescope (BAT) data were taken
period of 21.64±0.02 days. withtheHardX-rayTransientMonitorfrom16thDecember
Grebenevet al haverecentlycalculated aphotonindex 2004 to the 31st May 2010 over 14-195 keV. A 58 month
of0.5±0.5usingdatafromINTEGRAL.Theyhavealsocre- light curve was downloaded from NASA’s Swift/BAT 58-
atedenergyspectraover3-200keVshowingthattheenergy Month Hard X-ray Survey4. This contained an average of
ofthehighenergycut-offinthespectrumisat16.0±5.0keV ∼15 observations a day split into 8 energy bands.
(Grebenev,Lutovinov and Tsygankov 2012).
InthispaperwelookatSwiftJ045106.8-694803inmore
2.1.3 RXTE
detail, confirming its luminosity, pulse period, orbital pe-
riod and spin-up rate which can be used to determine the Archival data were taken from HEASARC. These were
strength of its magnetic field. recorded in two datasets on the 28th October 2011 using
the Rossi X-ray Timing Explorer’s (RXTE) Proportional
CounterArray,over ∼3-10 keV,as summarised in Table 2.
2 OBSERVATIONS Cleaned light curves were produced for each dataset
andcombined.ALomb-Scarglenormalisedperiodogramwas
2.1 X-ray Observations then created using a frequency intervalof 1×10-5 Hz.
2.1.1 Swift/XRT
Swift’s X-raytelescope (XRT)isaCCD imaging spectrom- 2.2 Optical observations
eter operating over 0.2-10 keV in photon counting mode. 2.2.1 Optical photometry
Archival data were taken from NASA’s High Energy As-
trophysics Science Archive Research Center (HEASARC)1 Archival data were taken from the MAssive Compact Halo
covering six time periods from 23rd October 2008 to 3rd Objects(MACHO)projectusingthe1.27metertelescopeat
October 2011, as summarised in Table 1. the Mount Stromlo Observatory in Australia. This covered
The images were extracted using the ftool2 xselect. the period from 1st November 1992 to the 29th December
Source and background spectra were then extracted from 1999 and contains instrumental magnitudes using red (R)
regions of 34(cid:2)(cid:2)radii. The spectra were binned to have 50 and blue(B) filters.
counts per bin. The ancillary response files (ARF) were The data were filtered to remove results flagged as er-
calculated with xrtmkarf and a redistribution matrix file roneous. The four points remaining in the R band dataset
(RMF) was taken from HEASARC’s calibration database whichwereover2.4standarddeviationsfromthemeanwere
(CALDB). The position of the source was confirmed using also removed.Thisleft 96datapointsinthered and207in
ftool xrtcentroid. theblue.ALomb-Scarglenormalisedperiodogramwasthen
The total count rate and error of each dataset, as createdinPeriod forboththeRandBbanddatasetsusing
1 http://heasarc.gsfc.nasa.gov/ 3 http://www.starlink.rl.ac.uk/docs/sun167.htx/sun167.html
2 http://heasarc.nasa.gov/ftools/ 4 http://swift.gsfc.nasa.gov/docs/swift/results/bs58mon/
(cid:3)c 2012RAS,MNRAS000,1–12
Swift J045106.8-694803; a highly magnetised neutron star in the Large Magellanic Cloud. 3
a frequency interval of 1×10-5 Hz and detrending using a
polynomial of order 3.
2.2.2 Optical spectroscopy
Optical spectra of Swift J045106.8-694803 have been taken
on three separate occasions. Red end spectra were taken
with the 1.9m Radcliffe telescope at the South African As-
tronomicalObservatory(SAAO)onthe12thDecember2009
and26thSeptember2011.Thedatawereobtainedusingthe
unitspectrographcombinedwitha1200linesmm−1grating
andtheSITedetectorattheCassegrainfocus.Theresulting
spectrahaveaspectralresolutionof∼3˚A.Comparisoncop-
per neon spectra were taken immediately before and after
theobservation and were used for wavelength calibration. Figure 1.UncertaintycirclesoftheX-raysourceoverlaidonan
The source was also observed on the 8th and 10th De- opticalbackground(fromESODSSIIBlue).Theredcircleshows
thepositionfoundusingtheSwift/XRTdataset00038029001,the
cember 2011 with the New Technology Telescope (NTT),
blue circle using dataset 00038029002 and the green using the
La Silla, Chile. Grisms 14 and 20 on the ESO Faint Ob-
XMM-NewtondatasetfromBartlettet al (2012, inprep).
ject Spectrograph (EFOSC2) were used for blue and red
end spectroscopy respectively with a slit width of 1.5(cid:2)(cid:2).
Grism 14 has a grating of 600 lines mm−1 that yielded
1 ˚A pixel−1 dispersion over a wavelength range of λλ3095–
5085˚A.Grism20isoneofthetwonewVolume-PhaseHolo-
graphic grisms recently added to EFOSC2. It has a smaller
wavelengthrange,from6047–7147 ˚A,butasuperiordisper-
sion of 0.55 ˚A pixel−1 and 1070 lines pixel−1. FilterOG530
wasusedtoblocksecondordereffects.Theresultingspectra
were recorded on a Loral/Lesser, thinned, AR-coated, UV
flooded, MPP CCD with 2048×2048 pixels, at a spectral
resolution of ∼ 10 ˚A and ∼ 6 ˚A respectively. Wavelength
calibration was achieved using comparison spectra of He-
liumandArgonlampstakenthroughouttheobservingrun
with thesame instrument configuration.
The data were reduced using the standard packages
available in the Image Reduction and Analysis Facility
(IRAF). The resulting spectra were normalized to remove
the continuum and and a redshift correction applied corre-
Figure 2. The upper panel shows the X-ray spectrum from
spondingtotherecessionvelocityoftheLMC,takenas-280
km s−1 (Paturel et al 2002). Swift/XRT dataset 00038029001 over 0.2-10 keV, here the pho-
tonindexis0.7±0.1withaNH of3±2counts/s,thelowerpanel
showsresiduals.
3 RESULTS
Figure 4 shows the Lomb-Scargle normalised peri-
3.1 X-ray Results odogram for each of the six datasets. A period is only ev-
ident in datasets 00038029001 and 2, these are 187.0±0.3
3.1.1 Swift/XRT
and186.8±0.2secondsrespectively,andshownoharmonics.
Figure 1 shows the positions calculated for datasets Thisindicatesthatlightcurvesfoldedatthesevaluesshould
00038029001(red)and2(blue).ThefirstdatasethasanRA, besinusoidal.Foldedlightcurvesfordatasets00038029001-2
Dec (J2000) of 04:51:06.4 and -69:48:02.5 with a two sigma are shown in Figure 5.
error radius of 3.6(cid:2)(cid:2).The second has an RA,Dec (J2000) of Lightcurvesfordataset00038029001ineachofthefour
04:51:07.0 and -69:48:03.1 with a two sigma error radius of energyrangesdeterminedfromFigure2areshowninFigure
3.6(cid:2)(cid:2). These are consistent with the positions calculated by 6.Aplotofpulse-fractionagainstenergyisshowninFigure
Beardmore et al (2009). 7.Thepulse-fractionappearstoincreasewithincreasingen-
The results of data extracted using xspec are sum- ergy between 0.5 and 4.5 keV and then decrease, however
marised in Table 3. Figure 2 shows the X-ray spectra of thelarge error bars make thisinconclusive.
dataset 00038029001 over 0.2-10 keV. This shows that the The side peaks in the periodograms are due to gaps in
countscanbedividedequallyintofourenergyrangesof0.5- theobservations.Thiswasconfirmedfirstly,byfittingasine
1.5, 1.5-3, 3-4.5 and 4.5-8 keV. The X-ray luminosities are wave with the same period to the data, which gave rise to
plotted in Figure 3, which shows that the X-ray luminosi- identical peaks. Secondly, by splitting dataset 00038029001
tiesmeasuredbetweenOctober2008andJanuary2009were into6shorterdatasetscomposedofcontinuousobservations
aboutfourtimeshigherthanthosemeasuredinOctoberand and adding the individual periodograms, as can be seen in
September2011. Figure 8.
(cid:3)c 2012RAS,MNRAS000,1–12
4 H. Klus, E.S. Bartlett, A.J. Bird, M. Coe, R.H.D. Corbet and A. Udalski
ObservationID CountRate NH Photon Flux Luminosity Reduced Degreesof
(counts/s (1021 cm-2) Index (10-11 ergs/ (1036 ergs/s) Chi-Squared freedom
over0.2-10keV) s/cm2)
a 00038029001 0.198±0.006 3±2 0.7±0.1 2.3±0.1 7.0±0.3 1.4 21
b 00038029002 0.175±0.005 1±2 0.7±0.1 2.1±0.1 6.3±0.4 1.2 17
c 00038029003 0.182±0.008 3±7 0.6±0.2 2.3±0.2 7.0±0.6 0.6 5
d 00038029004 0.181±0.006 1±3 0.6±0.1 2.1±0.1 6.5±0.4 1.2 11
e 00038029005 0.073±0.004 15±14 1.1±0.4 0.7±0.2 2.2±0.5 0.8 2
f 00038029006 0.058±0.003 1±16 0.6±0.4 0.8±0.2 2.3±0.5 2.0 2
Table3.SummaryofinformationextractedfromthespectraofSwift/XRTdatasets00038029001-6over0.2-10keVusingadistanceof
50.6±1.6kpc(Bonanos et al 2011).
Figure 3. Luminosity plotted against time for the Swift/XRT Figure 5. Light curves from Swift/XRT datasets 00038029001
and XMM-Newton datasets over 0.2-10 keV. XMM-Newton (top panel) and 00038029002 (bottom panel) over 0.2-10 keV,
datasetfromBartlettet al (2012, inprep). foldedat187.0and186.8secondsrespectively.Thephaseshiftis
arbitrary.
Figure4.Lomb-ScarglenormalisedperiodogramsforSwift/XRT
datasets00038029001-6(labeleda-f)over0.2-10keV.Thedotted
Figure 6. Light curves from Swift/XRT dataset 00038029001,
linesindicatethe187secondperiodandthetwosidepeaksof181
folded at 187.0 seconds, over 0.5-1.5, 1.5-3, 3-4.5 and 4.5-8 keV
and193secondsmentionedbyBeardmoreet al (2009).
(labelleda-d).
Datasets 00038029001 and 2 provided the best signal
and so their light curves were combined to determine the Bootstrapping was conducted in order to confirm that this
best estimation for the pulse period in late 2008. A Lomb- is the correct peak, and not a product of the window func-
Scargle normalised periodogram yielded a maximum peak tion,andafter5000iterationsbetween166and200seconds,
at 186.85±0.04 seconds as can be seen in Figure 9. Monte aperiod of187.07±0.04 seconds wasfound 63% of thetime
Carlo simulations give this period a 99.9% confidence level. and181.41±0.04seconds37%ofthetime.Wethereforecon-
(cid:3)c 2012RAS,MNRAS000,1–12
Swift J045106.8-694803; a highly magnetised neutron star in the Large Magellanic Cloud. 5
Figure 7. Pulse-fraction plotted against energy for Swift/XRT
dataset 00038029001 foldedat 187.0seconds, over 0.5-1.5, 1.5-3,
3-4.5and4.5-8keV.
Figure 9. Lomb-Scargle normalised periodogram from
Swift/XRT datasets 00038029001 and 2 over 0.2-10 keV
showing a maximum peak at 186.85±0.04 seconds. The side
peaksareduetogapsintheobservations.
for theBATdataset usingPeriod with a frequencyinterval
of 1×10-5 Hzbut no periods were detected.
3.1.3 RXTE
TheLomb-Scarglenormalisedperiodogramisshowninplots
(a) and (b) of Figure 11. As with the Swift/XRT data, the
side peaks are due to gaps in the observations. This was
confirmedbyfittingasinewavewiththesameperiodtothe
Figure 8. Lomb-Scargle normalised periodogram created by data,ascanbeseeninplot(c)ofFigure11.Theharmonics
splittingdataset00038029001into6datasetscomposedofcontin- at exactly 1/3 and 1/4 of the pulse period, seen in plot
uousobservationsandaddingtheperiodograms.Thedottedline (a), indicate that the light curve should be non-sinusoidal
indicatesthe187.0secondperiodfoundfordataset00038029001. andhavemultiplepeakswhenfoldedatthepulseperiod,as
shown in Figure 12.
cludedthat187.07±0.04 secondswasthemostprobablepe- The shape of the light curve contains information on
theemissiongeometryfromtheregionsclosetotheneutron
riod of Swift J045106.8-694803 in October/November 2008.
stars magnetic poles. At high luminosities photons can es-
cape from the sides of the accretion column, giving rise to
3.1.2 Swift/BAT and INTEGRAL/IBIS a fan-beam pattern and a more complex profile than that
foundatlowerluminositieswhereX-raysaregenerallyemit-
ThetoppanelofFigure10showsthelong-termlightcurveof
ted in a pencil-beam.
Swift J045106.8-694803 with total counts over 14-195 keV,
The RXTE data show a pulsation period of 169.8±0.3
binned at 28 days. The count rate appears to increase over
seconds, 17.3±0.3 seconds less than the period evident in
time, peaking in about July 2007 and continuing to remain
the Swift/XRT datasets taken in October and November
above zero for the nextthree years.
2008. This gives a spin-up rate (P˙) of -5.8±0.1 s/yr, or
Thiswasconfirmedbythelightcurvecompiledfromthe
INTEGRAL IBIS data which is shown in the second panel (-1.84±0.03)×10-7 s/s, and a spin-up time scale (Ts) of -
30.8±0.5 yrusing
ofFigure10.Thebottompanelshowsthelightcurveofthe
optical counterpart taken from OGLE III and IV in the I- P
band.The OGLE data shows that thesource was brightest Ts= −P˙ (1)
in the I-band whilst it was barely detectable in the X-ray.
ThereisaslightincreaseinbrightnessafterMJD54000but whereP istheaveragepulseperiodinseconds.Sincethisis
this is not significant. an average over three years, it is possible that the spin-up
A Lomb-Scargle normalised periodogram was created rate has been much higher at pointsduring this period.
(cid:3)c 2012RAS,MNRAS000,1–12
6 H. Klus, E.S. Bartlett, A.J. Bird, M. Coe, R.H.D. Corbet and A. Udalski
Figure 10.Long-termlightcurvefromSwift/BAT(14-195keV),INTEGRALIBIS(15keV-10MeV)andOGLEIIIandIV.
Figure 11. Panel (a) shows the Lomb-Scargle normalised peri- Figure 12. Light curve folded at 169.78 seconds from RXTE
odogramfromthecombinedRXTEdatasetsover3-10keV.Panel datasets 96441-01-01-00and01over3-10keV.
(b)showsacloseupviewoftheregionaroundthemainpeakcom-
paredtopanel (c) whichshowsthe resultofasimulateddataset
producedbyapuresinewaveofperiod169.78seconds. anup-to-dateperiodandluminosity.Thepositionwasfound
to be at an RA, Dec (J2000) of 04:51:06.7 and -69:48:04.2
respectively, with a one sigma uncertainty of 1(cid:2)(cid:2). The lumi-
3.1.4 XMM-Newton
nositywas found tobe(9.8±0.9)×1034 ergs/s - asshown in
Swift J045106.8-694803 was observed by the X-ray Multi- Figure 3 - and theperiod 168.5±0.2 seconds.
MirrorMission-Newton(XMM-Newton)on17thJuly2012 This is 1.3±0.4 seconds less than the period calculated
(Bartlett et al 2012, in prep).Thisconfirmedthepositionof from RXTE in 2011, giving a P˙ of -1.8±0.5 s/yr. It is
Swift J045106.8-694803 -asisshowninFigure1-andgave also 18.6±0.2 seconds less than the period calculated from
(cid:3)c 2012RAS,MNRAS000,1–12
Swift J045106.8-694803; a highly magnetised neutron star in the Large Magellanic Cloud. 7
Figure 13. Plot showing the three periods measured by Figure 15. Lomb-Scargle normalised periodogram for the B
Swift/XRT in 2008, RXTE in 2011 and XMM-Newton in 2012 band MACHO dataset showing a possible orbital period of
(Bartlett et al, 2012, in prep). The straight line is a line of best 21.631±0.005days.
fitshowingthattheperiodiscontinuingtodecrease, despitethe
factthattherateofchangeisalsodecreasing.
Figure16.LightcurvefromtheBbandMACHOdatasetfolded
at21.631days.
Figure14.LightcurvesfortheRandBbandMACHOdatasets.
Swift/XRTin 2008, giving an average P˙ of -5.01±0.06 s/yr form the Be stars circumstellar disk. If the rotational ve-
andaTs of-34.9±0.4 yr.Figure13showsthethreeperiods locity of the Be star is close to the critical velocity, then
measuredbySwift/XRTin2008,RXTEin2011andXMM- pulsationsleadtomatterbeingejectedandspunuptoform
Newton in 2012. The luminosity and rate of change of spin a Keplerian disk (Reig 2011).
periodmaybedecreasingbutSwiftJ045106.8-694803 isstill This period is not evident in the R band, this is most
continuingto spin-up at a high rate. likely due to the lack of data points rather than the conse-
quenceofarealeffect.Thisideawasconfirmedbyrandomly
removinghalfoftheBbanddatapointsandcreatinganew
3.2 Optical results Lomb-Scargle normalised periodogram, which also failed to
showanyevidenceofanorbitalperiod.Alightcurvefolded
3.2.1 Optical photometry
at21.631dayswasthenproducedfortheBbanddatasetas
Figure 14 shows that the flux of the optical companion to showninFigure16.Colourratioandcolourmagnitudedia-
SwiftJ045106.8-694803appearstohaveremainedfairlycon- gramswerealsoproduced.Theseconfirmedthatthemagni-
sistent in the B and R bands over 7 years. The B band tudeineachcolourbandhasremainedfairlyconsistentover
datasetshowapossibleorbitalperiodof21.631±0.005 days theobservation period.
asseeninFigure15,althoughanyunderlyingnon-radialpul- Swift J045106.8-694803 was also observed as part
sations from the Be star may affect the results (Bird et al of the OGLE III (LMC136.6.14874) and OGLE IV
2012). Non-radial pulsations occur when some parts of the (LMC531.05.4251) programmes. These I-band data cover,
stellar surface moveinwards while othersmoveoutwardsat in total, a duration of more than a decade, and are shown
the same time. It has been suggested that they could help inthebottompanelofFigure10.Fromthisfigureitisclear
(cid:3)c 2012RAS,MNRAS000,1–12
8 H. Klus, E.S. Bartlett, A.J. Bird, M. Coe, R.H.D. Corbet and A. Udalski
that the source undergoes a significant long-term modula- theabsolutemagnitudeofemission linestars,which willbe
tion on periods in excess of 400 days. Formally, a Lomb- dependentontheinclinationandsizeofthedisc.Assuchwe
Scargle power spectrum gives the peak to be around 440 classifytheopticalcounterpartofSwiftJ045106.8-694803 as
days, but a visual inspection of the light curve shows that a B0-1 III-Vstar.
there are other time-scale changes occurring. It is therefore Figure 18 shows the three red end spectra of Swift
unlikelythatthislongperiodisrelateddirectlytothebinary J045106.8-694803 taken two years apart. The ESO (top)
period, but rather may either indicate a general time-scale spectrumisoffsetby6fluxunits.TheHαequivalentwidth,
for fluctuations in the stellar wind, or precessional motion consideredanindicatorforcircumstellardisksize,isremark-
of the circumstellar disk. All these variations make it very ably similar in all three spectra increasing from -29±2 ˚A
difficult to search for confirmation of the 21.631 day pe- for the SAAO spectrum taken in 2009 to -33±1 ˚A and -
riod seen in theMACHOdata. However,if onlythebetter- 34.5±0.6 ˚A for the SAAO and ESO spectra taken in 2011.
sampledOGLEIVdataaremergedwiththeMACHOdata Itisnot uncommontoseelarge variationsintheHαequiv-
(normalised to the approximate starting magnitude of the alent width of these systems on timescales of months. This
OGLE III data), then the strength of the 21.631 day peak apparentconsistencyintheequivalentwidth(andhencethe
in theLomb-Scargle power spectrum increases slightly. But disksize)isalmostcertainlylinkedtotheexceptionallyper-
without the prior knowledge of this period from the MA- sistent X-rayactivity of thesource.
CHO data such a period would not have been found in the
OGLE data.
3.2.2 Optical spectroscopy
4 DISCUSSION AND CONCLUSIONS
OBstarsintheMilkyWayareclassifiedusingcertainmetal
We have confirmed the pulsation and orbital periods given
and helium line ratios (Walborn and Fitzpatrick 1990) byBeardmoreetal (2009).Recentmeasurementsshowthat
based on the Morgan-Keenan system (MK; Morgan et al SwiftJ045106.8-694803 hasspun-upatarateof∼5seconds
1943). However, this is unsuitable in lower metallicity en-
a year for the last four years and although this rate is de-
vironmentsasthemetallinesareeithermuchweakerornot
creasing with luminosity, it is still continuing to spin-up at
present. As such, the optical spectrum of IGR J05414-6858
a high rate.
wasclassifiedusingthemethoddevelopedbyLennon(1997)
Swift J045106.8-694803 may have experienced one of
forB-typestarsintheSMCandimplementedfortheSMC,
thehighestspin-upratesofanyknownaccretingpulsarbut
LMC and Galaxy byEvans et al (2004, 2006).
thiscanbeaccountedforusingGhoshandLamb’saccretion
Figure 17 shows the unsmoothed optical spectrum of
theory (1979) assuming that the stellar magnetic field is
Swift J045106.8-694803. The spectrum is dominated by the
dipolar and that it is accreting from a Keplerian disk.
hydrogenBalmerseriesandneutralheliumlines.Theredoes
Accretion theory shows that the magnetic field (B) of
appeartobeevidencefortheHeiiλ4200˚Aline,butitisdif-
a disc-fed accreting pulsar is related to the rate of change
ficulttodistinguishabovethenoiselevelalongwiththeHeii of its spin period (P˙), its average spin period (P) and its
λλ4541,4686˚Alines.TheHeiλ4143˚Alineisclearlystronger
average luminosity (L).
than theHeiiλ4200˚A lineconstraining theoptical counter- Ingeneral, thehigherthemass accretion rate(M˙),the
part of Swift J045106.8-694803 to be later than type O9.
higher the luminosity and the more angular momentum is
TherealsoappearstobeevidencefortheSiivλ4088,λ4116˚A
accretedpersecond.Wheremagneticstressesdominatemat-
lines necessary for a B1 classification. This is supported by
terflowatsurfaceS1–asseeninFigure19fromGhoshand
the relative strengths of the Siiiiλ4553˚A and Mgii λ4481˚A
Lamb (1979) - this almost always causes spin-up.
suggesting a classification of B2 or earlier.
AtsurfaceS2,whereviscousstressesdominate,thiscan
The luminosity class of the companion star was de-
add a spin-up or spin-down torque depending on the stars
termined by the ratios of Siv λ4088/Hei λλ4026−4121˚A,
Siv λ4116/Hei λ4121 and Heii λ4686/Hei λ4713. The first fastnessparameter(ωs).Thisistheratiobetweentheangu-
lar velocity of the neutron star and the Keplerian angular
two ratios strengthen with decreasing luminosity class (i.e.
velocity of thedisc.
with increasing luminosity) whereas the latter ratio de-
If the pulse period is long, plasma is able to penetrate
creases with increasing luminosity. The relative strengths
intothe magnetosphere and reach the neutron star surface,
of these lines are contradictory: The Heii λ4686/Hei λ4713
transferring angular momentum and causing a strong spin-
and Siv λ4116/Hei λ4121 ratios suggest a luminosity class
uptorque.Ifthepulseperiod isshort,theplasmaisunable
III, although the proximity of the rotationally broadened
to penetratefurther and theneutron star spinsdown.
Hδ λ4102˚A to the Siiv lines makes this more complex. The The magnetic moment (μ) can be found if P˙, P and L
Siv λ4088/Hei λ4026 ratio is more consistent with a star
are known byassuming a given mass and radiusand seeing
of luminosity class V. The V band magnitude of this star which value of μ predicts theobserved P˙ using
is reported by several sources as between 14.6 and 14.7
w(ea.gs.cMaalcsuselayte2d002u;siZnagritaskdyisettanacle20m04o)d.uAlunsMoVf 1o8f.-542.±3±0.00.71 −P˙ =5.0×10−5μ320/7n(ωs)S1(M)(PL337/7)2 (2)
(Bonanos et al 2011) along with an mV of 14.65±0.05 and where μ30 is the magnetic moment in units of 1030 Gauss
an extinction AV=0.4±0.1 (calculated using the Galactic cm-3 andL37 istheluminosityoftheaccretingstarinunits
column density towards the source, 8.4±1.0×1020 cm−2 of 1037 ergs/s. For an 0<ωs<0.9
and the results of Gu¨ver and O¨zel 2009). This is consistent
withB0.5III(Wegner2006).Lessinformationisavailablefor n(ωs)=1.4(1−2.86ωs)(1−ωs)−1 (3)
(cid:3)c 2012RAS,MNRAS000,1–12
Swift J045106.8-694803; a highly magnetised neutron star in the Large Magellanic Cloud. 9
1.4
Swift J045106.8−694803
1.3
1
2
1
4
λ
eI
1.2 6 & H
ελH3970 λHeI+II402 λSiIV4088δλH4102λSiIV4116 λHeI4143 λHeII4200 γλH4341 λHeI4387 λHeI4471 λHeII4541λSiIII4553 λCIII4650 λHeII4686λHeI4713
1.1
1.0
0.9
1
6
8
4
λ
β
H
0.8
4000 4200 4400 4600 4800 5000
Wavelength (Å)
Figure17.SpectrumofSwiftJ045106.8-694803inthewavelengthrangeλλ3900–5000˚AwiththeNTTon2011-12-08.Thespectrumhas
been normalized to remove the continuum andredshiftcorrected by -280km s−1. Atomic transitions relevant to spectral classification
havebeenmarked.
whereM˙17isthemassaccretionrateinunitsof1017grams/s
Dec 2011 and MM(cid:3) isassumedthroughouttobe1.4.S1(M)andS2(M)
are structure functions that depends on the mass equation
15
of state and thedynamical response of theneutron star.
d Flux 10 Sept 2011 S1(M)=R66/7(MM(cid:3))−3/7I4−51 (5)
e
s
mali whereR6istheradiusoftheaccretingstarinunitsof106cm,
or assumedthroughouttobe1,andI45isthemomentofinertia
N 5 Dec 2009 in units of 1045 grams cm2.
M
S2(M)=R6−3/7(M )−2/7 (6)
0 (cid:3)
6450 6500 6550 6600 6650 6700 Thevalueofthemagneticfieldcanthenbecalculatedusing
Wavelength (Å)
μ
B= (7)
R3
Figure 18. ESO (top panel) and SAAO (middle and bottom
panels)spectraofSwiftJ045106.8-694803inthewavelengthrange Sincen(ωs)dependsonlyonωs andωs isinverselypro-
λλ6400–6700˚A with the NTT on 2011-12-10. Spectra have been portional to PL373/7, the P˙ for a star of a given mass and
normalizedtoremovethecontinuumandshiftedby-280kms−1 magnetic moment depends only on (PL373/7)-1. Figure 20
shows that Swift J045106.8-694803 is located in the same
region as GX 1+4, SAX J2103.5+4545, 4U 2206+54 and
within 5% accuracy and SXP1062 on a plot of log10(P˙) against log10(PL373/7).
M GX 1+4 is part of a LMXB, it is a slow rotator
ωs=1.35μ63/07S2(M)(PL33/77)−1 =1.19P−1M˙1−73/7μ360/7(M )−5/7 with a period of about 121 seconds and has been ob-
(cid:3)
(4) served to have spun-down by about 2.6 seconds a year
(cid:3)c 2012RAS,MNRAS000,1–12
10 H. Klus, E.S. Bartlett, A.J. Bird, M. Coe, R.H.D. Corbet and A. Udalski
Figure 19.SideviewofaccretionflowfromGhoshandLamb(1979).
(Chakrabarty et al 1997). SAX J2103.5+4545 is part of a Orbitalperiod 21.631±0.005days
HMXBsystem,itisalsoaslowrotatorwithapulseperiodof PulseperiodinOctober 187.07±0.04s
358.166±0.0005secondsandaluminosityof(2.0±0.5)×1036 /November 2008
PulseperiodinOctober2011 169.8±0.3s
ergs/s. It has also been known to spin-up at a rate of
4.69±0.09 seconds per year (Sidoli et al 2005). This means PulseperiodinJuly2012 168.5±0.2s
AverageLuminosity (3.4±0.3)×1036 ergs/s
that it has a magnetic field of (1.47±00..0098)×1014 Gauss. P˙ -5.01±0.06s/yr
4U 2206+54 (Reig, Torrejon and Blay 2012) and SXP1062 (-1.59±0.02)×10-7 s/s
(Turolla and Popov 2012) are also part of HMXB systems, Ts -34.9±0.4yr
bothareknowntobespinningdownandbothwererecently log10(P˙) 0.70±0.01
identified as having extremely high magnetic fields. log10(PL373/7) 2.04±0.04
Highly magnetised neutron stars may be formed if a ωs 0.7±00..13
very magnetic star conserves its magnetic field during the n(ωs) -3±21
μ (1.2±0.2)×1032 Gausscm-3
neutron stars formation or, if the neutron star is rotating 0.7
fast enough to start a dynamo effect during the first ∼20 B (1.2±00..72)×1014 Gauss
seconds of its life. This converts heat and rotational en- Table 4.Asummaryofobserveddataandresultsderivedusing
ergy into magnetic energy and increases the magnetic field equations 1-7.forSwiftJ045106.8-694803.
(Ferrario and Wickramasinghe 2006).
Thefirstevidenceforhighlymagnetised accretingneu-
HMXB system containing LSI 61303 recently showed mag-
tron stars came from Pizzolato et al (2008), who showed
netar like behaviour when it underwent two bursts similar
that the X-ray source 1E161348-5055 may contain a neu-
tothoseofsoftgammarepeaters(SGRs)(Torres et al 2012;
tron star with a surface magnetic field of ∼1015G which is
Papitto, Torres and Rea 2012).
part of a LMXB. This was shortly followed by Bozzo et al
In conclusion, the spin-up rate of Swift J045106.8-
(2008), who argued that highly magnetised neutron stars
694803 can be explained using Ghosh and Lamb’s (1979)
may also exist in HMXB.
accretion theory provided that it has a magnetic field of
Our results show that the P˙ measured between 2008 (1.2±00..72)×1014 Gauss.
and2011 implies amagneticfieldof (1.3±0.1)×1014 Gauss,
whilstthatmeasuredbetween2011and2012impliesamag-
netic field of (6±1)×1013 Gauss. The value calculated for
thewholedurationofobservationsis(1.2±0.1)×1014 Gauss. ACKNOWLEDGEMENTS
Takingintoaccountthehighestandlowestpossiblevaluesof We acknowledge the use of public data from the Swift and
Bmeasuredbetween2008and2011and2011and2012,this RXTE data archive and are grateful for the advice from
valuebecomes(1.2±00..72)×1014Gauss.Parameterscalculated Gerry Skinner on the Swift/BAT data. This paper uti-
using equations 1-7 are shown in Table 4. lizes publicdomain data obtained by theMACHO Project,
All values are over the quantum critical value of jointly funded by the US Department of Energy through
4.4×1013 Gauss, which means that the physics must be the University of California, Lawrence Livermore National
described with quantum field theory rather than classical Laboratory undercontract No. W-7405-Eng-48, by the Na-
physics.AlthoughSwiftJ045106.8-694803 ispoweredbyac- tional Science Foundation through the Center for Particle
cretion, ratherthanduetothedecayand instabilities of its Astrophysics of the University of California under cooper-
magnetic field, and hence cannot be referred to as a mag- ative agreement AST-8809616, and by the Mount Stromlo
netar yet, it has the potential to become one. The Galactic and Siding Spring Observatory, part of the Australian Na-
(cid:3)c 2012RAS,MNRAS000,1–12