Timing and Spectral properties of Be/X-ray pulsar EXO 2030+375 during a Type I outburst

0.57 MB·English

by  Sachindra Naik

Draftversion December 11,2013
PreprinttypesetusingLATEXstyleemulateapjv.5/2/11
TIMING AND SPECTRAL PROPERTIESOF BE/X-RAY PULSAR EXO 2030+375DURING A TYPE I
OUTBURST
Sachindra Naik1, Chandreyee Maitra2, Gaurava K. Jaisawal1, Biswajit Paul2
1 Astronomy&AstrophysicsDivision,PhysicalResearchLaboratory, Ahmedabad-380009,Indiaand
2 RamanResearchInstitute, Sadashivnagar,C.V.RamanAvenue, Bangalore-560080, India
Draft version December11, 2013
ABSTRACT
3
We present results from a study of broadband timing and spectral properties of EXO 2030+375
1
using a Suzaku observation. Pulsations with a period of 41.41 s and strong energy dependent pulse
0
profiles were clearly detected up to 100 keV. Narrow dips are seen in the profiles up to ∼70 keV.
2
Presence of prominent dips at several phases in the profiles up to such high energy ranges were not

n seen before. At higher energies, these dips gradually disappeared and the profile appeared single-
a peaked. The 1.0-200.0 keV broad-band spectrum is found to be well described by a partial covering
J
highenergycut-offpower-lawmodel. Severallowenergyemissionlinesarealsodetectedinthe pulsar

7 spectrum. We fitted the spectrum using neutral as well as partially ionized absorbers along with
abovecontinuummodelyieldingsimilarparametervalues. Thepartialcoveringwithpartiallyionized
]
absorberresultedinto marginallybetter fit. The spectralfitting didnotrequireanycyclotronfeature
E
in the best fit model. To investigate the changesin spectralparametersat dips, we carriedout pulse-
H
phase-resolvedspectroscopy. Duringthedips,thevalueofadditionalcolumndensitywasestimatedto
. behighcomparedtootherpulsephases. Whileusingpartiallyionizedabsorber,thevalueofionization
h
p parameter is also higher at the dips. This may be the reason for the presence of dips up to higher
- energies. Nootherspectralparametersshowanysystematicvariationwithpulse phasesofthe pulsar.
o Subject headings: stars: neutron, pulsars: individual: EXO 2030+375,X-rays: stars
r
t
s
a 1. INTRODUCTION onds to several hundred seconds. The X-ray spectra of
[
Be/X-raybinariesrepresentthelargestsubclassofhigh these pulsars are generally hard. Fluorescent iron emis-
1 massX-raybinarysystems. Thecompactobjectinthese sion line at 6.4 keV is observed in the spectrum of most
v systems is generally a neutron star (pulsar) whereas the of the accretion powered X-ray pulsars. For a brief re-
2 companion is a B or O-type star which shows Balmer view on the properties of Be/X-ray binary pulsars, refer
1 to Paul & Naik (2011).
emission lines in its spectra. The binary optical com-
1 The transient X-ray pulsar EXO 2030+375 was dis-
panion lies well within the Roche lobe. The objects in
1 coveredduring a giant outburst in 1985 with EXOSAT
these binary systems are typically in a wide orbit with
. observatory (Parmar et al. 1989a). Optical and
1 moderate eccentricity. Though evolutionary model cal-
near-infrared observations of the EXOSAT error cir-
0 culationsshowthatbinarysystemswithwhitedwarfand
cle identified a B0 Ve star as the counterpart of
3 Be star or black hole andBe star shouldalso exist, clear
1 evidence of the existence of such binary systems has not EXO 2030+375 (Motch & Janot-Pacheco 1987; Coe et
: been found as yet (Zhang, Li & Wang 2004 and refer- al. 1988). Using the EXOSAT observations in 1985,
v thespinandorbitalperiodsofthe pulsarwereestimated
ences therein). The neutron star in these Be/X-ray bi-
i
X nary systems accrets matter while passing through the to be 42 s and 44.3-48.6 days, respectively. Analyz-
ing BATSE monitoring data of severalconsecutive out-
r circumstellardiskofthe companionBestar. Theabrupt
a accretion of matter onto the neutron star while passing bursts of the pulsar EXO 2030+375 in 1992, Stollberg
et al. (1997) derived following orbital parameters of the
through the circumstellar disk of the Be companion or
duringtheperiastronpassageresultsinstrongX-rayout- binary system : orbital period Porb = 46.02±0.01 days,
bursts (Okazaki & Negueruela 2001). During such out- e = 0.36±0.02, axsin i = 261±14 lt-sec, ω = 223◦.5 ±
1◦.8, and time of periastron passage τ = 2448936.8 ±
bursts, the X-ray emission from the pulsar can be tran-
siently enhanced by a factor more than ∼10. Be/X-ray 0.3 days. During the giant outburst in 1985, the pulsar
was observed with EXOSAT observatory. A significant
binary systems generally show periodic normal (type I)
change in 1-20 keV luminosity by a factor of ≥2500 was
X-ray outbursts that coincide with the periastron pas-
detected compared to that during the quiescent phase.
sage of the neutron star and giant (type II) X-ray out-
A dramatic change in pulse period was seen during the
bursts which do not show any clear orbital dependence
apart from the persistent low luminosity X-ray emission luminosity decline with spin-up timescale of −P/P˙ ∼
during quiescent (Negueruela et al. 1998). The neutron 30 yr (Parmar et al. 1989a). During the outburst, the
stars in the Be/X-ray binary systems are found to be pulseprofileofthe pulsarwasfoundtobe stronglylumi-
accretion powered X-ray pulsars except a very few cases nosity dependent. At high luminosity, the pulse profile
suchasLSI+61303(Massietal. 2004). Thespinperiod consisted of one main pulse and a small inter-pulse, sep-
of these pulsars is found to be in the range of a few sec- arated by ∼180◦ phase. The strength of the two pulses
was reversed when the luminosity was decreased by a
[email protected] factor of ∼100 (Parmar et al. 1989b). By using a ge-
2 Naik et al.
0.1
ometric model, Parmar et al. (1989b) explained that
the dominant beam of emission changed from a fan-
beamtoapencil-beamduringthedecreaseinluminosity
and that resulted in the strength reversal of the main 2007 May 14
)
and inter-pulse. An extensive monitoring campaign of V
e
EXO 2030+375 with BATSE and Rossi X-ray Tim- k
ing Explorer (RXTE) showed that a normal outburst 0
5
hasbeendetectedfornearlyeveryperiastronpassagefor −
∼13.5 years (Wilson, Fabregat & Coburn 2005). 15 0.05
The spectral analysis of EXO 2030+375 had been (1
−
carried out by using EXOSAT data during outburst s
2
(Reynolds, Parmar & White (1993); Sun et al. (1994)). −
m
The 1-20 keV spectrum was described by a two compo- c
nentcontinuummodelconsistingofa blackbodycompo- s
t
nentwithtemperature∼1.1keVwhereasthehardX-ray un
partwasrepresentedbyapower-law. RXTEmonitoring Co
of the pulsar during an outburst in 1996 June-July also
0
suggested that a two component (blackbody and power-
law with an exponential cut-off) model was required to 5.415×104 5.42×104 5.425×104 5.43×104
describe the 2.7–30 keV pulsar spectrum (Reig & Coe Time (MJD)
1999). A spectral feature at ∼36 keV in the hard X-ray
spectrum (in 17-65 keV range) was ascribed to a pos- Fig.1.—TheSwift/BATlightcurveofEXO2030+375 in15-50
keVenergyband,from2007February4(MJD54135)to2007July
sible cyclotron absorption line implying the estimated
29 (MJD 54310). The arrow mark shows the date of the Suzaku
magnetic field of the pulsar to be 3.1×1012 Gauss (Reig observationofthepulsar.
& Coe 1999). However, using regular monitoring data
of EXO 2030+375 with the RXTE from 2006 June to databetween2007February4and2007July29covering
2007 May, covering the first giant outburst since its dis- thepresenttype-Ioutburst. Thearrowmarkinthefigure
coveryin1985,Wilsonetal. (2008)reportedacyclotron shows the Suzaku observation of the pulsar during the
feature at ∼11 keV and estimated the magnetic field peakoftheoutburst. Thisobservationwascarriedoutat
strength to be 1.3×1012 Gauss. This feature was con- “HXDnominal”pointingpositionforeffectiveexposures
sistently detected in the pulsar spectrum for about 90 of ∼57 ks and ∼53 ks for XIS and HXD, respectively.
days when the 2-100 keV luminosity was above 5×1037 The XISwere operatedwith “burst”clockmode in “1/4
erg s−1. INTEGRAL and Swift observation of the window” option, covering 17′.8×4′.4 field of view.
same giant outburst was used to describe the 3-120 keV The fifth Japanese X-ray astronomy satellite Suzaku
spectra by using an absorbed power-law with an expo- (Mitsuda et al. 2007) was launched on 2005 July 10.
nential cut-off, an iron emission line and some peculiar There are two sets of instruments onboardSuzaku such
featuresin10-20keVenergyrange. The featurein10-20 as X-ray Imaging Spectrometer (XIS; Koyama et al.
keV energy range was modeled by a broad emission line 2007) which covers 0.2-12 keV energy range and Hard
at ∼13-15 keV or by two absorption lines at ∼10 keV X-rayDetector(HXD;Takahashietal. 2007)whichcov-
and ∼ 20 keV (Klochkov et al. 2007). Pulse-phase re- ers 10-600 keV energy range. Among the four sets of
solved spectroscopy of INTEGRAL observation of the XIS, each with a 1024×1024 pixel CCD at the focus of
pulsarduringthegiantoutburstshowedsignificantspec- correspondingtelescope, one(XIS-1)is back-illuminated
tralvariabilityofthecontinuumparameters(Klochkovet whereas the others are front illuminated. In full window
al. 2008). In the process, evidence of the presence of an mode, the field of view of the XIS is 18’×18’ with an
absorption line at ∼63 keV was found at a narrow pulse effective area of 340 cm2 and 390 cm2 at 1.5 keV for
phase interval when EXO 2030+375 was at the peak of front-illuminated and back-illuminated CCDs, respec-
its giant outburst. This feature was interpreted as the tively. The non-imaging instrument HXD which was de-
harmonic of the previously reported ∼36 keV cyclotron signed to detect high energy X-ray photons, consists of
line. 16identicalunitsmadeupoftwotypesofdetectorssuch
For a detailed study of timing andspectralproperties, as silicon PIN diodes covering 10–70 keV energy range
EXO 2030+375 was observed with Suzaku on 14 May andGSO crystalscintillatorcovering30–600keVenergy
2007, at the peak of a regular Type I outburst. The range. Theeffectiveareaofthe PINdiodesis∼145cm2
resultsobtainedfromthe timingandspectralanalysisof at 15 keV whereas that of GSO detectors is ∼315 cm2
the Suzaku observation are presented in this paper. at 100 keV. As XIS-2 was unoperational during the ob-
servationof EXO 2030+375,data from other 3 XISs are
2. OBSERVATION used in the present analysis.
Observation of the transient Be/X-ray binary pulsar
EXO 2030+375 was carried out on 2007 May 14 at the 3. ANALYSISANDRESULTS
peakofitsregulartypeIoutburst. Thepubliclyavailable For XIS and HXD data reduction, we reprocessed the
arcivaldata(ver-2.0.6.13)ofaboveobservationwasused unfiltered eventdata using ’aepipeline’ packageof HEA-
in the present work to investigate the properties of the Soft version 6.12 and utilizing the calibration database
pulsar during the outburst. Figure 1 shows the one-day (CALDB) released on 2012 February 10 (for XIS) and
averaged light curve of EXO 2030+375 in 15-50 keV 2011 September 13 (for HXD). Source light curves and
energy range obtained from the Swift/BAT monitoring spectra were extracted from the reprocessed cleaned
Suzaku Observation of EXO 2030+375 3
1.5 1.5
XIS−0 0.5−2.0 keV
1
1
1.5
0.5 2−4 keV
y
sity 1.5 PIN nsit 1
n e
e t
t n
n I
d I 1 ed 1.5
e z 4−6 keV
z i
li al
a m
m 0.5 1
r
r o
o 1.5 N
N
GSO
1.5
6−10 keV
1
1
0.5
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Pulse Phase Pulse Phase
Fig.2.— XIS-0 (top panel; in 0.2–12 keV range), PIN (second Fig.3.—TheXIS-0pulseprofilesofEXO2030+375 atdifferent
panel;in10–70keVrange)andGSO(bottompanel;in40–600keV energy ranges. Thepresence of several dip likefeatures in0.3-0.7
range)pulseprofiles,obtainedfromthecorrespondinglightcurves and 0.9-1.1 pulse phase ranges are clearly seen. The error bars
by using the estimated 41.4106 s pulse period of EXO 2030+375. represent1σ uncertainties. Twopulsesareshownforclarity.
Theerrorsinthefigureareestimatedforthe1σ confidence level.
Twopulsesareshownforclarity.
(in 0.2–12keV energy range),PIN (in 10–70keV energy
event data of XIS, PIN and GSO detectors. The simu- range), and GSO (in 40–600 keV energy range) event
latedbackgroundevents,assuggestedbytheinstrument data, respectively. Pulse folding and a χ2 maximization
team, were used to estimate the PIN and GSO back- technique was applied to the light curves obtained from
grounds for the EXO 2030+375 observation. The re- XIS,PINandGSOeventdataandthepulseperiodofthe
sponse file which was releasedin 2008January was used pulsarwasestimatedtobe41.4106(1)s. Thelightcurves
for HXD/PIN spectrum whereas response and effective were then folded with this period to get the pulse pro-
areafilesreleasedin2010MaywereusedforHXD/GSO. files at different energy range. It is found that the pulse
The reprocessed XIS event data were checked for the profilesofthe pulsarobtainedfromthe backgroundsub-
possible presence of photon pile-up. Pile-up estimation tracted light curves of XIS-0, XIS-1 and XIS-3 are iden-
was performed by examining the Point Spread Function tical, whereas it is somewhatdifferent to the profiles ob-
(PSF) of the three XISs by checking the count rate per tained from HXD/PIN and HXD/GSO data. The pulse
oneCCDexposureattheimagepeakasgivenbyYamada profilesofEXO2030+375 obtainedfromthebackground
& Takahashi1. It was found that the XIS event data subtractedXIS-0,HXD/PINandHXD/GSOlightcurves
was not affected by photon pile-up. The source spec- of the Suzaku observation are shown in Figure 2. The
tra were accumulated from the XIS reprocessed cleaned XIS and HXD/PIN profiles, though look similar, the
′
eventdata byselecting acircularregionof3 aroundthe structure and depths of the dips are found to be dif-
image centroid. The XIS backgroundspectra were accu- ferent. However, the difference is very clearly visible in
mulated from the same observation by selecting circular the HXD/GSO profile. To investigatethe energy depen-
regions away from the source. The response files and dence of the pulse profile of EXO 2030+375, we gener-
effective area files for XIS were generated by using the ated source and corresponding background light curves
”xissimarfgen” and ”xisrmfgen” tasks of FTOOLS. in various energy bands from the XIS, PIN and GSO
event data. After appropriate background subtraction,
3.1. Timing Analysis the lightcurveswerefoldedwiththe estimatedpulse pe-
For the timing analysis, the arrival times of the X-ray riod and the corresponding pulse profiles are shown in
photons were converted to the same at the solar system Figures 3 & 4. From the figures, it can be seen that the
barycenter by using the Suzaku specific barycenter cor- X-ray pulsations in EXO 2030+375 are clearly seen up
rection task “aebarycen”. X-ray light curves with time to ∼100 keV. The dips in the profiles are very strong
resolutions of 2 s, 1 s and 1 s were extracted from XIS and clearly distinguishable up to ∼40 keV. Beyond this
energy, the width and depth of the dips in the pulse
1http://www-utheal.phys.s.u-tokyo.ac.jp/yuasa/wiki/index.php/Howprofiles decrease gradually up to ∼70 keV. Beyond ∼70
tocheckpileupofSuzakuXISdata keV,however,thedipsbecomeindistinguishableandthe
4 Naik et al.
pulse profiles appear smooth and single peaked. The
presence of dips in the pulse profiles up to ∼70 keV is
rarely seen in the accretion powered X-ray pulsars. The N(E)=e−σ(E)NH1(S1+S2e−σ(E)NH2)E−ΓI(E)
characteristicpropertiesofthesedipsinthe pulseprofile
where
of EXO 2030+375 is being investigated in the present
work through pulse phase resolved spectroscopy. I(E)=1 for E <E
c
− E−Ec
3.2. Spectral Analysis =e (cid:16) Ef (cid:17) for E >Ec
3.2.1. Pulse phase averaged spectroscopy NH1 is the Galactic equivalent hydrogen column den-
We analyzed the pulse-phase-averaged spectra of sity, NH2 is the additional equivalent hydrogen column
EXO 2030+375 using XIS, HXD/PIN and HXD/GSO density of the material local to the neutron star, N(E)
eventdata. The spectra fromboth the front-illuminated is the observed intensity, Γ is the photon index, σ is the
CCDs(XIS-0andXIS-3)wereaddedtogetheralongwith photo-electriccross-section,S1andS2arethenormaliza-
correspondingresponsematricesandbackgroundspectra tionsofthepower-law,Ecisthecut-offenergyandEf the
by using package “addascaspec”. Data from XIS-1 was e-folding energy. The covering fraction of the absorbed
used separately in the spectral fitting. In the spectral power-lawdue to the presence ofadditionalmatter local
fitting,weselecteddataintheenergyrangesof1-10keV to the neutron star is expressed as Norm2 / (Norm1 +
for both front and back-illuminated CCDs (added spec- Norm2) = S2/(S1+S2)].
tra from XIS-0 and XIS-3, and XIS-1), 12-70 keV for As mentioned earlier, the relative instrument normal-
the HXD/PIN and40-200keVfor the HXD/GSO. After izations of added front-illuminated CCDs (XIS-0 and
appropriate backgroundsubtraction, simultaneous spec- XIS-3, quoted as XIS03), XIS-1, PIN and GSO detec-
tralfitting was carriedout using the XIS, PIN andGSO tors were kept free. The corresponding values obtained
spectra with XSPEC v12.7.1. All the spectral param- are 1.00:1.04:0.99:1.02 for XIS03:XIS1:PIN:GSO with a
eters other than the relative instrument normalization, clear agreement with the values at the time of detector
were tied together for all the detectors. The XIS spec- calibration. After fitting the spectra with above model,
tra were binned by a factor of 5 from 1-10 keV whereas excess residuals were found to be present at ∼2 keV,
the PIN spectrum was not binned up to 50 keV beyond ∼2.5 keV, ∼3.2 keV and ∼6.6 keV. As the observation
whichitwasbinnedbyafactorof3. TheGSOspectrum wasduring the peak ofthe Type I outburstand the pul-
was binned with the fixed grouping scheme provided by sar being bright at hard X-rays, the presence of several
the instrument team2. Because an artificial structure is emission lines is expected in the spectrum. Therefore,
knownto be presentin the XIS spectra at aroundthe Si we added Gaussian functions at above energies to the
edge, we ignored data between 1.7–1.9 keV and 2.2–2.4 spectral model and re-fitted the spectra. Though these
keVinourspectralanalysis. Inthebeginning,wetriedto emission lines are weak, addition of these lines to the
fitthebroad-bandspectraofthepulsarwithvariouscon- model improved the simultaneous spectral fitting with
tinuum models such as power-law model modified with reduced χ2 of 1.59 (for 638 dof). Detection of several
an exponential cut-off, Negative and Positive power law emission lines at soft X-rays and the pulsar being so
withEXponential(NPEX)continuummodelandpartial bright during the Suzaku observation, we attempted to
coveringpower-lawwithhighenergycut-offmodelalong fitthe broad-bandspectrawiththe partialcoveringhigh
withinterstellarabsorptionandIronKα emissionlineat energycut-offpower-lawmodelwithpartiallyionizedab-
6.4 keV.It is found that in case of all continuummodels sorber(zxipcf modelinXSPEC;Reevesetal. 2008)and
otherthanthepartialcoveringhighenergycut-offmodel, Gaussian functions for detected emission lines. Using
the spectral fitting was extremely poor with reduced χ2 the partiallyionized abosrbercomponentin the spectral
of more than 3. Therefore, we ignored all other contin- model, the ionization states of the absorbing medium
uummodelsinourfitting. Thoughafewofthesemodels andthecorrespondingcoveringfractioncouldbeinvesti-
were used to describe the source continuum earlier (for gated. The zxipcf model component in XSPEC uses a
example - Reynolds et al. 1993; Wilson et al. 2008), grid of XSTAR (Kallman et al. 2004) photoionized ab-
highersensitivitydetectorswithbetterenergyresolution sorption models for absorption. The free parameters in
onboard Suzaku and its broad-band spectral capability thismodelareNH (in1022 cm−2),C (coveringfraction),
provide good statistical quality in the spectrum to rule andtheionizationparameterξ (ergcms−1;Reevesetal.
out the other spectral models for this source. It may 2008). Thespectralfittingwasimprovedmarginallywith
happen that the properties of the pulsar may be differ- reduced χ2 of 1.51 (for 637 dof). The energy spectra of
ent during the normal outburst (present one) and giant thepulsarEXO2030+375 areshowninFigure5&6(for
outbursts (Reynolds et al. 1993; Wilson et al. 2008). partialcoveringmodelswithneutralandpartiallyionized
absorbers, respectively) along with the spectral compo-
3.2.2. Partial covering high energy cutoff power-law model nents (top panels) and residuals to the best-fit model
(bottom panels). As the difference in both the spectral
The partial covering power-law with high energy cut-
modelsisthenatureoftheadditionalabsorptioncompo-
offmodelconsistsoftwopower-lawcontinuawithacom-
nents without any noticable change in the best-fit value
monphotonindexbutwithdifferentabsorbinghydrogen
of other spectral parameters, both the figures look sim-
column densities. The analytical form of the partially
ilar. The best-fit parameters obtained from the simul-
covering power-law with high energy cut-off model is
taneous spectral fitting to the XIS, PIN and GSO data
with partial covering model modified with neutral and
2http://heasarc.gsfc.nasa.gov/docs/suzaku/analysis/gsobgd64bins.datpartially ionized absorbers are given in Table 1.
Suzaku Observation of EXO 2030+375 5
1.4
10−20 keV PIN 40−55 keV GSO
1.5
1.2
1
1
0.5 0.8
1.5 20−30 keV PIN 55−70 keV GSO
1.5
1
1
0.5
y
t 2
nsi 1.5 30−40 keV PIN 70−100 keV GSO
e 1.5
t
n
I 1 1

d
e
z 0.5
i 0.5
l
a
m 3
100−200 keV GSO
Nor 1.5 40−55 keV PIN 2
1
1
0
0.5
−1
55−70 keV PIN 200−600 keV GSO
2 1.5
1
1
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Pulse Phase Pulse Phase
Fig.4.—TheenergyresolvedpulseprofilesofEXO2030+375 obtainedfromHXD/PINandHXD/GSO lightcurves atvariousenergy
bands. The presence/absence of dip like structures can be seen in 0.3-0.7 and 0.9-1.1 pulse phase ranges. The error bars represent 1 σ
uncertainties. Twopulsesineachpanelareshownforclarity.
The spectral fitting of the RXTE observations of the 10
pulsarduring2006June giantoutburstshowedthe pres- V−1
enceofacyclotronresonancefeaturecenteredat∼11keV e 1
k
(kWeVilswoanserteaplo.r2te0d08a)t. Acerctyacilnotpruonlserepsohnaasencreanligneesaotf∼th63e nts s −1 0.1
u
pulsar from the INTEGRAL observations during same o
C 0.01
giant outburst (Klochkov et al. 2008). From RXTE d
e
observations of the pulsar at relatively lower luminosity aliz 10−3
level, a spectral feature was detected at ∼36 keV and m
or
ascribed to a cyclotron absorption feature (Reig & Coe N 10−4
4
1999). However, in our spectral fitting, no such absorp- 2
tion feature was present at above energies. Therefore, χ 0
−2
we did not include any additional cyclotron absorption −4
component to the spectral fitting. 10 100
Energy (keV)
3.2.3. Pulse phase resolved spectroscopy Fig. 5.— Energy spectra of EXO 2030+375 obtained with the
XIS, PIN and GSO detectors of the Suzaku observation at the
The presence of several prominent dips in the pulse peak of the type I outburst, along with the best-fit model con-
profile and partial covering power-law with high energy sisting of a partially absorbed power-law with high energy cutoff
cut-off being the best fit continuum model imply the continuum model with neutral absorber. Five emission lines (as
given inTable 1) aredetected inthe broad-band spectrum of the
possibilities of presence of streams of matter at various
pulsar. Thecontributionsoftheresidualstotheχ2foreachenergy
pulse phases around the poles of the transient Be/X-
binforthebest-fitmodelareshowninthebottom panel.
ray binary pulsar EXO 2030+375. To investigate this,
6 Naik et al.
10
TABLE 1
Best-fit parametersof the phase-averagedspectra for V−1
EXO 2030+375 during 2007MaySuzakuobservation with e 1
k
aGb1saσoursebsriearrno)rchsoi.gmMhpooendneeenlr-tg1sy,:cMPuaotrdotefilfa-l2poc:woPveaerrr-tliniaawgl(mcwooivdteherlninweguitt(hrwafitlivhe nts s −1 0.1
u
partiallyionized absorber) highenergycutoffpower-law Co 0.01
modelwith fiveGaussiancomponents. d
e
z
Parameter Value ali 10−3
m
Model-1 Model-2 or
NH1 2.04±0.01 2.07±0.01 N 10−44
NH2 40.9±3.3 45.4±2.8 2
CoveringFraction 0.13±0.01 0.23±0.01 χ 0
−2
log(ξ) —— 1.56±0.15 −4
Power-lawindex 1.33±0.01 1.37±0.01 10 100
Highenergycutoff(keV) 12.4±0.2 11.0±0.8 Energy (keV)
E-foldenergy(keV) 21.2±0.2 21.5±0.2
Fig.6.—EnergyspectrumofEXO2030+375 obtainedwiththe
XIS,PINandGSOdetectorsoftheSuzakuobservationatthepeak
Emissionlines of the type I outburst, along with the best-fit model comprising
apartiallyabsorbedpower-lawwithhighenergycutoffcontinuum
SiXIII modelwithpartiallyionizedabsorber. Theemissionlines(asgiven
Lineenergy(keV) 2.01±0.01 2.01±0.01 in Table 1) detected in the pulsar spectrum are represented with
Linewidth(keV) 0.001 0.001 Gaussianfunctionsinthefigure. Thecontributionsoftheresiduals
Lineeq. width(eV) 2 2±2 to the χ2 for each energy bin for the best-fit model are shown in
thebottom panel.
SiXIV energy band to each of the 25 phase resolved XIS and
Lineenergy(keV) 2.50±0.03 2.51±0.02 PIN spectra. The phase resolved XIS and PIN spectra
Linewidth(keV) 0.13±0.03 0.14±0.02 werefittedwiththepartialcoveringpower-lawwithhigh
Lineeq. width(eV) 11 13 energy cut-off model with neutral and partially ionized
absorber (as was done in phase-averaged spectroscopy).
SXV
Some of the parameterssuchas the values ofrelative in-
Lineenergy(keV) 3.19±0.01 3.19±0.01
Linewidth(keV) 0.11±0.02 0.12±0.01 strumentnormalizations,theGalacticabsorption(NH1),
center energy and width of emission lines were fixed at
Lineeq. width(eV) 8 10
the values obtained from phase-averaged spectroscopy.
IronKα Theparametersobtainedfromthesimultaneousspectral
Lineenergy(keV) 6.39±0.01 6.4±0.02 fitting to the XIS and PIN phase resolved spectra using
Linewidth(keV) 0.01+−00..0021 0.05±0.02 partialcoveringmodelwithneutralandpartiallyionized
Lineeq. width(eV) 7 19 absorber are shown in the left and right panels of Fig-
ure7,respectively. Thetopthreepanelsinboththesides
FeXXVI ofFigure7showthepulseprofilesofEXO2030+375 ob-
Lineenergy(keV) 6.61±0.01 6.66±0.02 tained from XIS, PIN and GSO data. The values of ad-
Linewidth(keV) 0.20±0.02 0.08±0.03 ditionalcolumndensity(NH2)andcoveringfractionover
Lineeq. width(eV) 51 25
pulse phases are shown in fourth and fifth panels of the
figure, respectively. The change in values of power-law
Sourceflux
11700---107200k0ekVekVeVrarnarnagnegege 252...995423+−+−+−000000......010010735625 252...995424+−+−+−000000......010010854937 ppevlahusrolisrtaeeotsnpipohneiancstodeifevsxepa,loyrwe.e-esfTohr-lohldaweiwnnlgeinfnetonrssmeiixrdatgehlyi,zbaasoenttvidtoeonnmctuwhtph-aoaenffnrdeeaelesnisgehthrhogtewhysrpoigatvhhneetr-
Reducedχ2 1.59(638) 1.51(637) side bottom panel of the figure shows the value of ion-
izationparameterlog ξ overpulsephases(forthepartial
NH1 =Galacticequivalenthydrogencolumndensityalongthe
lineofsight,NH2 =Additionalhydrogencolumndensityof covering model with partially ionized absorber).
materiallocaltotheneutronstar. NH1 andNH2 areintheunits Amongthespectralparameters,mostnotableandsys-
of(1022 atomscm−2). Sourcefluxisintheunitof10−9 ergs
tematic variabilityis seeninthe values ofadditionalcol-
cm−2 s−1. Sourcefluxquotedaboveisnotcorrectedfor
umn density and covering fraction of the absorber. The
interstellarabsorption.
values of NH2 are found to be high at phases where dips
pulse-phase-resolvedspectroscopy was performed on the are present in the pulse profiles. This pattern is seen in
XIS and PIN data. By applying phase filtering in the case of both the models i.e. partial covering model with
FTOOLS task “XSELECT”, we accumulated 25 pulse neutral as well as partially ionized absorber. In case of
phase resolved spectra from XIS and HXD/PIN event partial covering model with partially ionized absorber,
data. Data from GSO detectors were not used in phase- the value of the ionization parameter was found to be
resolved spectroscopy as the signal-to-noise ratio was relativelyhigheratdip phases indicating the presence of
very poor for spectra of each phase bins. For phase- highlyionizedadditionalabsorberatseveralpulsephases
resolved spectroscopy, we used same background spec- causing dips in the pulse profiles. The higher values of
traandresponsematricesfor correspondingdetectorsas cut-off energy were also found at pulse phases that are
were used for phase-averaged spectroscopy. Simultane- coincident with the dips in the pulse profiles. The val-
ous spectral fitting was carried out in the 1.0-70.0 keV ues of the power-law photon index and e-folding energy
Suzaku Observation of EXO 2030+375 7
arefoundtobe variableoverthepulsephasesofthepul- the presence of additional absorption component at
sar. However,itisdifficulttocorrelatethesevariabilities certain pulse phases that partially obscured the emitted
with the dips in the pulse profiles. We did not find any radiation giving rise to dips in the pulse profiles. The
variabilityinthe fluxofboththeironemissionlinesover additional absorption is understood to be taking place
the pulse phases of the pulsar. This suggests that the bymatterinthe accretionstreamsthatarephaselocked
matter emitting the iron fluorescence lines is probably with the neutron star.
distributed symmetrically around the pulsar. In accretionpoweredpulsars,pulse profiles at hardX-
rays are generally simpler and smoother than that at
4. DISCUSSION lowenergies. Low-energypulseprofilesaremoreaffected
by absorptionby circumstellarmatter and/oradditional
Be/X-ray binaries usually show two types of X-ray
matter distribution near the neutron star. In case of
outbursts such as normal (type I) outbursts and giant
majority of the Be/X-ray binary pulsars, the soft X-ray
(type II) outbursts. The normal outbursts are char-
acterized by lower luminosity (∼1036−37 erg s−1) and pulseprofilesarefoundtoconsistofdipsatvariouspulse
phaseswhereas the hardX-raypulse profiles aresmooth
occur near the periastron passage. However, the giant
outbursts are characterized by high luminosities (≥1037 andsingle-peaked. EXO2030+375 is oneofafew other
erg s−1) and very rare (Stella, White & Rosner 1986; accreting X-ray pulsar which clearly show hard X-ray
pulsations. In the present work, it was found that the
Negueruela et al. 1998; Bildsten et al. 1997). In case
41.41 s pulsations are detected in the X-ray light curves
ofBe/X-raypulsarEXO2030+375,the type I outbursts
at high energies (∼100 keV). Similar results are seen in
are seen almostevery ∼46 days (orbitalperiod) whereas
a few other pulsar such as 1A 1118-61(up to ∼100 keV;
the giant outbursts were only seen twice; in 1985 during
Coe et al. 1994; Bildsten et al. 1997), 2S 1417-624 (up
which the pulsar was discovered (Parmar et al. 1989a,
to ∼100 keV; Finger et al. 1996), GX 1+4 (up to ≥100
1989b) and in 2006 (Wilson et al. 2008). The pulsar
keV;Naik,Paul&Callanan2005),GROJ1008-57(upto
has been monitored extensively with various X-ray ob-
∼100keV;Naiketal. 2011)etc. Highenergyphotonsbe-
servatories giving rise to the accurate estimation of the
ing less affected by the absorption/scattering by matter
orbital parameters of the binary system (Wilson et al.
in the interstellar medium as well as matter distribution
2008). Attempts have been made to measure and un-
close to the neutron star, the hard X-ray pulse profiles
derstand the complex nature of the phase-averagedpul-
possiblyindicate theintrinsicradiationpatternfromthe
sar spectrum (Klochkov et al. 2008; Wilson et al. 2008
pulsar. Analyzingtheshapeofenergyresolvedpulsepro-
andreferencestherein). Phase-resolvedspectroscopyhas
files of EXO 2030+375 during its secondgiant outburst
been performedonly once when the pulsar was undergo-
in2006,Sasakietal. (2010)modeledthegeometryofthe
ing the second giant outburst in 2006 (Klochkov et al.
neutron star by identifying the emission components of
2008). Though,phaseresolvedspectralanalysiswascar-
the magnetic poles. The asymmetric shape of the pulse
riedoutonthe RXTE observationsofthe pulsarduring
profilesduringthegiantoutburstwasexplainedinterms
a normal outburst (Reig & Coe 1998), the inferior spec-
ofmoderatelydistortedmagneticfieldtheconsequenceof
tralcapability and low energy thresholdof the detectors
whichisthatonepoleofthepulsargetsclosertotheline
limited the understanding of the complex nature of the
ofsightthantheother. Becauseoftheasymmetryinpo-
pulse profile. Broad-bandcapabilityandhighsensitivity
sitionofthepoles,thesymmetricpulseprofilesfromboth
of detectors onboard Suzaku provided the opportunity
the poles merged together and appeared as asymmetric
to investigate the properties of the pulsar at different
inshape. Similar analysiswascarriedoutto understand
pulse phases in more detail.
the geometry of the neutron star in other binary pulsars
suchas4U0115+63andV0332+53atvariousluminosity
4.1. Pulse Profile
levels oftheir outbursts (Sasakiet al. 2012). In contrast
The pulse profile of the transient pulsar to the results obtained from the 2006 giant outburst of
EXO 2030+375 are found to be strongly energy EXO 2030+375 (Klochkov et al. 2008; Sasaki et al.
dependent. The presence of several prominent dips 2010), we find that at high energies, the shape of the
makes the soft X-ray pulse profiles complex. The pulse profiles appeared to be symmetrical. At soft X-
strength of these dips gradually decreases with energy, rays, the pulse profile was found to be complex because
making the profile a smooth and single peaked at high of the presence of several absorption dips. The shape of
energy bands. X-ray pulsations in EXO 2030+375 are the observed pulse profiles during type I outburst sug-
detected up to as high as ∼100 keV. The shape of the gests the asymmetric distribution of the magnetic poles
pulse profiles obtained from Suzaku observation are which was used to describe the findings during the giant
found to be different from that obtained from RXTE outburst in EXO 2030+375may not be applicable.
(Reig & Coe 1998;Sasakiet al. 2010), JEM −X (Mar-
4.2. Spectroscopy
tinez Nunez et al. 2003), INTEGRAL (Klochkov et al.
2008) observatories, though a few of these observations Thebroad-bandX-rayspectrumofEXO2030+375 in
were carried out during the 2006 giant outburst. The 1-200 keV energy range has been described here in this
presence of dips in the pulse profile are seen in other paper for the first time in detail. Camero Arranz et al.
transient X-ray pulsars such as A0535+262 (Naik et al. (2005) presented the pulsar spectrum in 5-300 keV en-
2008), GX 304-1 (Devasia et al. 2011), GRO J1008-57 ergy range obtained from the INTEGRAL observation
(Naik et al. 2011), 1A 1118-61 (Devasia et al. 2011; of a type I outburst in 2002 December by a model con-
Maitra et al. 2012), RX J0440.9+4431 (Usui et al. sisting of a disk blackbody (kT ∼ 8 keV) and a power-
2012), Vela X-1 (Maitra & Paul 2012) etc. In most of law with Γ of 2.04 or a Comptonized component. The
the cases, pulse-phase resolved spectral analysis showed lowerthresholdat5keVlimitedtheunderstandingofthe
8 Naik et al.
1.5 1.5
XIS XIS
1 1
y y
it 0.5 it 0.5
s s
en 1.5 PIN en 1.5 PIN
t t
n 1 n 1
I I
d 0.5 d 0.5
e e
liz 1.5 GSO liz 1.5 GSO
a a
m m
1 1
r r
o o
N N
150 150
NH2(10)22 10500 NH2(10)22 10500
g 0.6 g 0.6
nn nn
verictio00..24 verictio00..24
oa oa
Cfr Cfr
1.6 1.6
Γ 1.4 Γ 1.4
1.2 1.2
) 25 ) 25
EfoldkeV 1250 EfoldkeV 20
( (
20 15
) 15 ) 15
utV 10 utV 10
Ec(ke 5 Ec(ke 5
1 2
rm ξg 0
o 0.5 o −2
N l
−4
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Pulse Phase Pulse Phase
Fig.7.— Best-fit spectral parameters obtained from the phase resolved spectroscopy of Suzaku observation of EXO 2030+375. The
XIS(in0.2-12keVenergyrange),PIN(in10-70keVenergyrange)andGSO(in40-600keVenergyrange)pulseprofilesofthepulsarare
shownintopthreepanelsofbothsidesofthefigure,respectively. Theotherpanelsintheleftsideshowthespectralparametersobtained
from pulse phase-resolved spectroscopy by using partial covering model with neutral absorber whereas the panels in the rightside shows
thesameforpartialcoveringmodelwithpartiallyionizedabsorber. Theerrorsshownareestimatedfor1σ confidence level.
characteristic properties of the pulsar at soft X-rays. A gen column density increases significantly. This increase
blackbodycomponentatatemperatureof8keVisrarely in column density modifies the emitted radiation at soft
reported so far and difficult to explain in accretionpow- X-rays. While simultaneous spectral fitting to the pul-
ered X-ray pulsars. They also did not find any evidence sar spectrum in 1-200 keV energy range, the estimated
of presence of an iron line or cyclotron line features in valueofcolumndensityalwaysexceededthecorrespond-
the pulsar spectrum. RXTE observations of the pulsar ing Galactic value in the direction of the pulsar. Even
during type I outbursts, though used a blackbody and though, we attempted to fit the broad-band spectra of
power-law components to explain the 2.7-30 keV spec- EXO 2030+375 with various continuum models, partial
trum, the blackbody temperature was estimated to be covering model provided best-fit to the data. Though
≤1.4 during entire outburst (Reig & Coe 1999). The the estimated value of absorption column density was
presence of an iron emission line at 6.4 keV was seen higher, another significant absorption component with
in the RXTE spectra. Apart from the RXTE observa- certain value of covering fraction (as given in Table 1)
tions, there are other observations at high energy bands wasrequiredin the spectralfitting. EXO 2030+375 be-
some of which are during the two giant outbursts. ingabrighthardX-raypulsarandpresenceofsignificant
In case of Be/X-ray binaries, the circumstellar disk amountofmatteratperiastronpassage,there isa possi-
aroundthe Becompanionplaysanimportantroleinthe bility of ionizing the surrounding matter. To investigate
X-ray emission from the neutron star. During the peri- this, we attempted to fit the spectrum with an partially
astronpassage,theouteredgeofthediskistruncatedby ionizedabsorberwhichyieldedmarginalimprovementin
theneutronstarresultinginX-rayoutbursts. Duringthe thespectralfitting. Apartfromthis,lowenergyemission
periastronpassage,asadditionalmatterisbeingaccreted linesarealsodetectedinthe pulsarspectrum. We found
onto the neutronstar,the value ofthe equivalent hydro- SiXIII,SiXIV,SXV,FeKαandFeXXVIemissionlines
Suzaku Observation of EXO 2030+375 9
in the XIS spectra of the pulsar. Though the multiple profilesarefoundto be stronglyenergydependent. Nar-
iron emission lines are seen in other accretion powered row dips which are generally seen in the pulse profiles
X-ray pulsars e.g. Cen X-3 (Naik, Paul, & Ali 2011 and of accretion powered pulsars up to ∼10 keV, are seen
referencestherein),GX1+4(Naik,Paul&Callanan2005 in the profiles up to as high as 70 keV. At soft X-rays,
andreferencestherein),thelowenergySiandSemission the shape of the pulse profiles is found to be complex
lines are detected for the first time in this pulsar. As whereas at high energies, it appeared symmetrical. The
the previous observations were carried out during giant complex nature of the profiles at soft X-rays is inter-
outburstsorwiththeinstrumentsoflowerspectralcapa- preted as because of the presence of several absorption
bilityatsoftX-rays,theseSiandSlinesweremissedout dips. Apartialcoveringpower-lawwithhighenergycut-
fromdetection. Duringgiantoutburstswhicharenotas- off continuum (with neutral as well as partially ionized
sociated with the periastron passage, lack of significant absorber)model was found to be the preferred model to
amountofadditionalmatterinthecloseproximityofthe describe the broad-band spectrum in 1.0-200.0 keV en-
neutronstarasdetectedduringtypeIoutbursts,reduces ergy range. High values of additional column density at
the chance of presence/detection of these emission lines the dip phases in the pulse profile confirm the presence
inthespectrum. ThespectralfittingofSuzakuobserva- of stream of absorbers that are phase locked with the
tion of the pulsar did not show any evidence of presence pulsar. Apart from the fluroscence iron emission line,
ofcyclotronabsorptionfeatures atearlierreportedener- several low energy lines from S and Si are also detected
gies from other observations. in the pulsar spectrum. Cyclotron resonance scattering
Pulse-phase resolved spectroscopy of the Suzaku ob- features, though reported earlier in the spectrum of this
servationofEXO2030+375 showedthatthevalueofad- pulsar, are not detected in the 1.0-200.0 keV spectrum
ditional column density (NH2) due to the material local ofthepulsar. Pulse-phaseresolvedspectroscopyrevealed
to the neutron star was about two orders of magnitude thatthe highervalues ofionizationparameteratthe dip
higher at certain pulse phases. It can be seen (from the phases of the pulse profile may be the cause of the pres-
top panels of Figure 7) that absorption dips are present ence of dips up to ∼70 keV.
atsamepulsephaseranges. Theadditionalhighvalueof
NH2, therefore, explains the presence of absorption dips
in the pulse profile. Apart from the variation of the ad- ACKNOWLEDGMENTS
ditional column density and covering fraction over pulse We thank the referee for his/her suggestions that im-
phases, the value of high energy cut-off was found to be provedthe presentationofthe paper. The researchwork
higheratdipphases. However,otherparametersdidnot at Physical Research Laboratory is funded by the De-
show any systematic variation over the pulse phases. partment of Space, Government of India. The authors
would like to thank all the members of the Suzaku for
5. CONCLUSION theircontributionsintheinstrumentpreparation,space-
In this paper, we performed timing and broad-band craft operation, software development, and in-orbit in-
spectralanalysisontheSuzakuobservationoftheBe/X- strumental calibration. This research has made use of
ray transient pulsar EXO 2030+375 during the peak of data obtained through HEASARC Online Service, pro-
a type I outburst. The 41.41 s pulsations were detected vided by the NASA/GSFC, in support of NASA High
in the lightcurves up to as high as ∼100keV.The pulse Energy Astrophysics Programs.
REFERENCES
Angelini,L.,Stella,L.,Parmar,A.N.,1989,346,906 NaikS.,PaulB.,Ali,Z.,2011, ApJ,737,79
Bildsten,L.,etal.,1997,ApJS,113,367 Negueruela,I.,Reig,P.,Coe,M.J.,&Fabregat,J.,1998, A&A,
CameroArranz,A.Wilson,C.A.,Connell,P.,MartinezNunez, 336,251
S.,Blay,P.,Beckmann,V.,Reglero,V.,2005,A&A,441,261 Okazaki,A.T.,Negueruela, I.,2001,A&A,377,161
CoeM.J.,etal.,1994,A&A,289,784 Parmar,A.N.,White,N.E.,Stella,L.,Izzo, C.,Ferri,P.,1989a,
Coe,M.J.,Longmore,A.,Payne,B.J.,Hanson,C.G.,1988, ApJ,338,359
MNRAS,232,865 Parmar,A.N.,White,N.E.,Stella,L.,1989b, ApJ,338,373
Devasia,J.,James M.,PaulB.,IndulekhaK.,2011,MNRAS, Paul,B.,Naik,S.,2011,BASI,39,429
414,1023 Reeves,J.,Done,C.,Pounds,K.,etal.2008,MNRAS,385,L108
Devasia,J.,James M.,PaulB.,IndulekhaK.,2011,MNRAS, Reig,P.,Coe,M.J.,1998,MNRAS,294,118
417,348 Reig,P.,Coe,M.J.,1999,MNRAS,302,700
Finger,M.H.,Wilson,R.B.,Chakrabarty, D.,1996,A&AS,120, Reynolds,A.P.,Parmar,A.N.,White,N.E.,1993,ApJ,414,302
209 Sasaki,M.,Klochkov,D.,Kraus,U.,Caballero,I.,Santangelo,
Kallman,T.,Palmeri,P.,Bautista, M.A.,Mendoza,C.,& A.,2010,A&A,517,A8
Krolik,J.H.,2004,ApJS,155,675 Sasaki,M.,Muller,D.,Kraus,U.,Ferrigno,C.,Santangelo, A.,
Klochkov,D.,etal.,2007,A&A,464,L45 2012,A&A,540,A35
Klochkov,D.,Santangelo, A.,Staubert,R.,Ferrigno,C.,2008, Stella,L.,White,N.E.&Rosner,R.,1986,ApJ,208,669
A&A,491,833 Stollberg,M.T.,1997,PhDthesis,Univ.Alabama
Koyama,K.,etal.,2007, PASJ,59,S23 Sun,X.-J.,Li,T.-P.,Wu,M.,Cheng,L.-X.,1994,A&A,289,127
Maitra,C.,Paul,B.,Naik,S.,2012, MNRAS,420,2307 Takahashi,T.,etal.,2007, PASJ,59,S35
Maitra,C.,Paul,B.,2012, ApJ(inpress),arXiv1212.1538 Usui,R.,etal.,2012,PASJ,64,79
MartinezNunez,S.,Reig,P.,Blay,P.,Kretschmar,P.,Lund,N., Wilson,C.A.,Fabregat, J.,Coburn,W.,2005,ApJ,620,L99
Reglero,V.,2003, A&A,411,L411 Wilson,C.A.,Finger,M.H.,Camero-Arranz,A.,2008,ApJ,678,
Massi,M.,Ribo,M.,Paredes,J.M.,etal.,2004, A&AL,414,1 1263
Mitsuda,K.,etal.2007,PASJ,59,S1 Zhang,F.,Li,X.-D.,Wang,Z.-R.,2004, ApJ,603,663
Motch,C.,Janot-Pacheco, E.,1987, A&A,182,L55
Naik,S.,Paul,B.,Callanan,P.J.,2005,ApJ,618,866
Naik,S.,etal.,2008,ApJ,672,516
NaikS.,PaulB.,KachharaC.,VadawaleS.V.,2011, MNRAS,
413,241