Properties of

High Mass X-ray Binary Pulsars :

emphasis on the Stellar Wind
of the companion
Uddipan Mukherjee (T.I.F.R., India)
Advisor : Prof. Biswajit Paul (T.I.F.R. & R.R.I., India)
OutlineoftheTalk
X-ray Binaries : Pulsars & High Mass X-ray Binaries
Stellar Winds
Detectors & Analysis
Orbital Phase Spectroscopy of HMXBs
4U153852
GX3012
OAO1657415
VelaX1
Be/X-ray binary 3A 0535+262 in quiescence

Spectral studies in the high and low states
CenX3
2S0114+650

Future Work
What are X-ray binaries?
Most bright X-ray
sources are in binaries
Compact object is
sucking material from
the other star
Systems composed of 2 stars
The other star is
called the companion
Binaries
According to the nature of the companion

If M
comp
M
sun
Low Mass XRBs

If M
comp
> M
sun
High Mass X-ray Binaries
Classification of XRBs
Compact object is a NS with strong
magnetic field (10
12
G)
Matter can only move following the
magnetic field lines
Accumulation of plasma on the NS
is called Accretion
X-rays make way out the magnetic
field in a narrow beam
X-ray pulsars
The radiation is in pulses and the
NS is an X-ray pulsar
High Mass X-ray Binaries
HMXRB contain supergiants (O &B) as companion & a NS
(or a BH : Cyg X-1)
Companion is very bright (10
4
10
5

Lum of Sun)
Companion can be viewed in IR, optical or UV
NS can only be seen in X- rays
Most HMXRB are X-ray pulsars
As the supergiant expands, it increases in
brightness
The outer layers of its atmosphere push out
a stream of particles
NS can pick up some of these particles and
start accretion in X rays
What is a supergiant ?
Brightest & largest kind of stars
Radii : 20 to several hundred times of Sun
Red (Betelguese) Blue (Rigel)
Both types can explode as Supernovae
Red Giant
Star of at least 15 solar mass & exhausting H

He - C burning expands even larger

Red Supergiant
Emanates a vigorous Stellar Wind
Lose extended atmospheres

smaller, hotter Blue Supergiants


What are Stellar Winds ?
Stars emit not only radiation but also particles
The emission of particles is called the Stellar Wind
Continuous phenomena, not episodic outbursts
In a star like the sun, the wind arises from the corona
In hotter stars, the high radiative flux, drives the wind
Primarily by means of line scattering
2 important parameters of a Stellar Wind derived from
observations :
MassLossRate : M & TerminalVelocity : v

A star like Sun loses about 10
-14
solar mass yr
-1

from winds blowing with ~700 km s
-1
Hot luminous stars exhibit stronger winds blowing
at speeds up to ~2000 km s
-1
And loses up to 10
-5
solar mass yr
-1
via winds
They are important to know the mechanism of wind generation

Line Driven Winds

Winds of luminous hot stars are driven by absorption in spectral lines
Hot stars emit bulk of their radiation in UV
Radiative acceleration in the winds of hot stars
mainly by the absorption & re-emission of UV photons
resonance lines of ions of abundant elements
C, N, O & Fe-group elements
Winds from quasars may also be radiation driven
Resonance line of N IV ~ 765
N
+++
ion absorbing it increases its velocity by (hv/mc) ~37cm s
-1
To accelerate a single N
+++
ion to the terminal velocity of 2000 km s
-1
,
we need 5 10
6
absorptions
In a plasma, momentum shared with other constituents
Hence, on an average : 10
11
absorptions per ion to boost the velocity
to the terminal value
The terminal velocity of a wind is reached within a few stellar radii
Forexample
To solve the Eqn. Of Motion for a stationary,
time independent wind
Assumptions :
Photosphere as a point source : radiation in the radial
direction
A photon emitted from the photosphere is absorbed by a
line transition in the wind in only a narrow interaction
region
The wind is isothermal and behaves like a perfect gas
Then, the solution is :
v(r) = v

(1 R/r)
0.5

where R is the stellar radius
S
o
b
o
l
e
v

A
p
p
r
o
x
i
m
a
t
i
o
n
Velocity Profile of Line Driven Winds
CAK Wind Profile
Observations & Analysis

To go above the Earth's atmosphere

Possible via balloons, rockets & satellites

Data from ASCA, BeppoSAX, Chandra,

RXTE & XMM-Newton
X-rays from Stars
Satellite
Image Spectrum Light Curve
X-ray emission
mechanisms !!
Any
Periodicities ??
Position
RXTE PCA : Large Area (1000 sq. cm. @ 6 keV)
Broad Energy Band (2 60 keV)
Time Resolution : 1 s
Energy Res : 18% @ 6.0 keV

Averybriefdescription
ASCA : 0.4 10 keV
Energy Res : 2% @ 6 keV
Area : 100 sq. cm. @ 6 keV
BeppoSAX : 0.1 10 keV
Energy Resolution : 8% @ 6.0 keV
Area : 150 sq. cm. @ 6 keV
........Continued
Chandra :
ACIS : 0.1 10 keV

HETG : Energy Res : 0.5% @ 6 keV
Area : 30 sq. cm. @ 6 keV
XMM-Newton :
PIC : 0.1 15 keV
Area : 800 sq. cm. @ 6 keV (PN)
Energy Res : 2% @ 6 keV


Both have high earth orbits which helps to
provide long, continuous observations
A Glimpse of the Spectral Analysis
Assumeamodelfortheincidentspectrum
Generallyforpulsars:powerlaw(PL),blackbody(BB)&
highenergycutoffdescribetheircontinuum
Anyemissionlinesaremodeledasgaussianfeatures
Exponentiallineofsightabsorption
Incidentspectrum+detectorresponse=observedspectrum
trytomatchtheobservedspectrumbyvaryingtheparametersof
theassumedmodel

Line
of
sight
NS
Eccentric
orbit
Stellar
Wind
4rr
2
p(r)v(r) = M
-
N
H
= p(x)dx
X : path
Schematic diagram for HMXB pulsars
What we have done ............
EvaluatedtheColumnDensitiesbyanumericalintegrationalongtheline
ofsightfromtheNStotheobserveratinfinityastheNStraversedthe
ellipticalorbit
CAKvelocityprofilewasassumed
Theobserverwasonadiferentplanethantheorbit
Anassumedinclinationanglewastaken
Binaryparameters,Masslossrate&Terminalvelocitywereknown
Our Work
Part I
U. Mukherjee & B. Paul 2004, A&A,
427, 567
U. Mukherjee et al., 2006a
U. Mukherjee, B. Paul & S. Naik,
2006b
4U 1538-52 : a nice system with smooth outflowing wind
Properties
- Orbital Period ~ 3.73 d
- Eclipse duration ~ 0.6 d
- Eccentricity ~ 0.18
- Distance ~ 5.5 kpc
- X-ray Luminosity ~ 10
36
erg s
-1
- Companion Star : B-type supergiant (Parkes et al. 1978)
- Mass-loss rate ~ 10
-6
solar mass yr
-1
- v

: 1000 km s
-1
- Pulse Period ~529 s
Thus, it is a suitable candidate to study the wind
structure of the companion star
Observations : We had proposed to
observe the source with RXTE-PCA
P.I. : Prof. B. Paul

from 2003-07-31 to 2003-08-07

out-of-eclipse phases for 2 binary orbits
25 data segments : exposure of
(1.5 6.0) ks


Set III
- 1998-07-29 to 1998-08-01 with

BeppoSAX covering one binary
orbit
- 40 data segments : exposure of
1.7 4.0 ks
Set II
- 1997-01-01 to 1997-01-05 with
PCA
- 14 datasets : (1.5 10.0) ks
on-source time
.....supplementedwitharchivaldata
Spectral Analysis
Backgrnd sub source spectra in 3 20 keV
Model : P L + exp Cut-Off + line-of-sight
exp absp + gaussian @ 6.4 keV
All spectral parameters kept free for 2003
dataset
Line centre and width frozen for 1997
dataset
MECS and LECS data fitted in (1.8 10.0)
keV & (0.3 4.5 ) keV
No high-energy cut-off for SAX
The soft excess at 0.1 keV was too faint

2
:(0.8 1.5) : RXTE & (0.6 1.3) : SAX
LECS
MECS
RXTE-PCA
Spectral Parameters
I ~ (1.0 1.5), lack of any
binary phase modulation

Similarly for
E
c
: RXTE-PCA ~14 keV
& E
f
~7 keV
Slightly different from E
c
~ 16
keV &
E
f
~10 keV : BeppoSAX (Robba et
al. 2001)
Fluorescent Fe--line flux
measured with the average
spectrum taken over 2 -- 3 ks
Does not show any considerable
variation along the orbit
Variation of Column Density with orbital phase
Column density : smooth
variation over the orbital
phase
Model is consistent with
observations
2 HMXB pulsars showed
similar increase in column
density pattern near eclipse :
X1908+075 (Levine et al.
2004) & SMC X-1 (Woo et al.
1995)
They may also have isotropic
wind pattern from the
companion stars
4U 1538-52 has a moderate X-ray luminosity : 10
36
erg s
-1

No significant perturbation in the stellar wind acceleration
4U 1700-37 (Haberl et al. 1989), 4U 1907+09 (Roberts
et al. 2001) & GX 301-2 (Leahy 1991, Pravdo & Ghosh
2001) :
a simple spherical wind is not sufficient to explain the
column density profile
Conclusions for 4U 1538-52
- Continuum X-ray spectrum is hardly affected by the NS
revolution
- N
H
shows a smooth variation over the orbital phase
- A spherically symmetric stellar wind from the companion star may
produce the observed orbital dependence of N
H
for certain range of
the orbital inclination
- Orbital phase resolved N
H
measurements can be an
independent way to estimate the orbital inclination,
especially for non-eclipsing binaries
GX 301-2 : a system with clumpy wind structure
Orbital period : ~ 41.5 d
Eccentricity ~ 0.46
Distance ~ 5.3 kpc
X-ray Luminosity ~ 10
35--37
erg s
-1
Companion Star : B-type Supergiant (Parkes et al. 1980)
Mass-loss Rate ~ (3-10) 10
-6
solar mass yr
-1


v

~ 400 km s
-1
Pulse Period ~ 680 s
So, a large mass-loss rate & unusually low wind
velocity result in formation of clumped blobs of matter

Also, a large absorption column density is created
Thus, a variable luminosity & an eccentric orbit provide
a good site to probe the wind structure of the companion
Archival RXTE Observations
Set I : from 1996-10-05 to 1996-15-06 :

17 data segments
did not cover 0.85 - 0.98 phases
useful obs duration : 34 ks
Set II : from 2000-12--10 to 2000-19-11 :

39 data segments
almost a full phase coverage
useful obs duration : 262 ks
Choice of the spectral model

Her X-1 : Endo et al. (2000) : absorption had two components
One component absorbs the entire spectrum while the other
absorbs it partially
Spectrum fitted : Partial Covering Absorption Model (PCAM)
PCAM : 2 different power laws with same photon index but
different normalisations & absorbed by different column densities
Endo et al. (2002) & Saraswat et al. (1996) : ASCA spectra of
GX 301-2 with the PCAM + Fe K-alpha, K-beta emission lines and an
Fe absorption edge
The Chandra-HEG spectra of GX 301-2 was also fitted with the PCAM
A broad compton back-scattered peak @ 6.3 keV (660 eV
FWHM) in the Chandra spectrum (Mukherjee & Paul, 2003
& Watanabe et al. 2003)
For RXTE, we used the model used by Endo et al. (2002) + a
high energy exponential cut-off
The back-scattered peak was not included in the RXTE
spectrum due to the moderate energy resolution
Chandra
Wavelength (Ang)
C
o
u
n
t
s
/
s
/
A
n
g
6.3 keV
RXTE Spectral Results
Almost all datasets : good
Reduced
2
between 0.6 and 1.6
for 44 degrees of freedom (d.o.f.)
The fits with poor Reduced
2
showed wavy residuals with dips
around ~10, 20 and 30 keV
Not Clear if it is a systematic
phenomenon for this source
Systematic deviations at ~20 &
~40 keV observed with Beppo-
SAX also (Orlandini et al. 2000)
Poor
Good
10 spectral parameters varied : N
H1

N
H2
, I, the 2 normalizations, the 2
iron line intensities, edge depth, E
c

& E
f

5 kept frozen : the iron-line
energies & their FWHM along with
the edge energy (at the values of
Mukherjee & Paul 2003)
Photon Index
Cut-off Energy
e-folding energy
No remarkable change in the
continuum parameters
E
c
, E
f
& I are more or less
consistent with the previously
measured values (White et al.
1983; Orlandini et al. 2000).
The Continuum Parameters
Geometry of the GX 301-2/Wray
977 system (from Pravdo & Ghosh 2001)
Considerable increase in column density
near periastron; but also a substantial
scatter in the values at intermediate
phases
N
H1
N
H2
Large variation throughout the binary orbit
(from 10
22
to 10
24
H atoms cm
-2)
seems to be
one of the characteristics of the X-ray
spectrum of GX 301--2
The large variations of N
H1
& N
H2
at all
orbital phases indicate clumpiness of the
stellar wind at different size scales.
Peak between 0.1 and 0.2 was expected
since the line of sight passes through the
densest parts of the wind
Thus it is clear that the observed variation
in column density cannot be explained by a
spherically symmetric CAK wind only
Variation of the Column Densities
2 models put forward to explain the N
H
variation
Leahy (1991), from TENMA data, proposes a
spherically symmetric stellar wind + linear gas
stream
The peak near periastron was fitted when the
gas stream was introduced (see right, Leahy
1991)
A gas stream can be due to the dynamical
effect of the neutron star on the companion
wind and the effect would be strongest at
periastron
Comparison with the Models
Pravdo & Ghosh (2001), proposed the existence
of an equatorially enhanced circumstellar disk
about Wray 977
The model describes two interactions of the NS
with the disk which gives rise to the two peaks in
the column density
Model calculations are very different from the observed
variation with the RXTE-PCA data
There are probably strong inhomogeneties in the wind that
are causing large fluctuations in the column densities
Almost throughout the binary orbit the Covering Fraction
remains high
Indicates the presence of clumpy inhomogeneous material
A spherical wind plus + stream fitted the variation in column
density for 4U 1907+09 (Roberts et al. 2001)
Covering Fraction (CF)
defined as :
Norm2 / (Norm1+Norm2) ;
Norm1 and Norm2 are
respectively the normalizations
of the two power-laws
CF remains substantially
high almost throughout the
orbit which means that
there is dense and clumpy
material present throughout
Iron-line Flux & Equivalent Width
2 iron lines show large increases in flux near periastron and a possible small
increase near 0.1 (at least for 1996)
The peak near periastron (phase 0.9) is not very evident in the 1996 data due to
the lack of enough observations
Fluxes for both lines in the intermediate phases show a more or less steady value
EQW has a clear correlation with N
H2
, with considerable scatter
Makino et al. (1985) & Endo et al. (2002) have showed
that a correlation between N
H
& EQW exists in GX 301--2
Endo et al. (2002) discusses that the EQW increases
monotonoically with the column density as long as
the fluorescing plasma is optically thin and fully
surrounds the pulsar
The scatter seen in our correlation may be due to the
clumpiness of the reprocessing material
For small values of the N
H
, the EQW is constant
This can happen if the fluorescing material does not
surround the pulsar completely
Conclusions for GX 301-2
PCAM describes the X-ray spectrum well throughout the
binary orbit
Column Densities measured are very high with a large
variation indicating a clumpy nature of the stellar wind

Correlation of the Fe-line EQW with N
H2
suggests that
most of the Fe-line is produced by the local clumpy
matter
Both the models do not clearly explain our results
OAO 1657-415 & Vela X-1 : orbital phase dependent
spectroscopy
OAO 1657-415 Vela X-1

Orbital Period (d) : 10.44 8.96

Eclipse (d) : 1.7 2.0
eccentricity : 0.1 0.1

Distance (kpc) : 6.4 2.0
X-ray Luminosity (erg s
-1
) : ~10
37
410
36

Companion Star : B Supergiant B Supergiant
Pulse Period (s) : 38 283
Vela X-1 has fluctuations of luminosities to the order of 10%
These 2 comparable pulsars provide a good opportunity to
study their stellar winds through spectral variability
OAO 1657-415 :

RXTE archival data from
1997-10-31 to 1997-11-11 :
26 obs

On-source time : 1--18 ks
Fairly good coverage of the
orbit
Most purposes, pulsar was in
the eclipsed state
Also, we have analysed the
MECS obs on 2001-08-14 for
~104 ks

Vela X-1 :
RXTE archival data from 2005-
01-01 to 2005-01-09 : 38 obs

On-source time : 2--18 ks
PCU0 had lost the propane layer,
hence we did not use its data in
any further analysis
For a very high luminosity pulsar
like Vela~X-1, this is not a major
issue
The dataset does not provide a
comprehensive orbital coverage,
i.e. it lacks data for the 0.4--0.7
orbital phase duration
Observations
Quite similar to that used for GX 301-2
MECS spectrum of OAO 1657-415 best fitted with
PCAM + 2 gaussian lines, at ~6.5 keV (narrow) &
@ ~7.0 keV
Initial RXTE values of the parameters for the
iteration : From MECS fit
Similarly for Vela X-1 : Same model + only one
gaussian line worked best
RXTE is a non--imaging instrument
& OAO 1657-415 under the Galactic ridge
emission
So, explicitly incorporate the ridge background as
a separate spectral component : Raymond-Smith
plasma + power-law with appropriate
normalizations (Valinia & Marshall 1998)
For the RXTE spectra : systematics to the tune of
1%
Spectral Models
Continuum
OAO 1657-415 Vela X-1
OAO 1657-415
Vela X-1
C
o
v
e
r
i
n
g

F
r
a
c
t
i
o
n
C
o
v
e
r
i
n
g

F
r
a
c
t
i
o
n
Column
Density
&
Covering
Fraction
Any Conclusions ??
The PCAM model appears to be somewhat generic for
pulsars which have variable column density, at least
when observed with RXTE/SAX
For highly luminous pulsars, the CAK Model of stellar
wind does not suffice to describe the column density
variations
Moderately luminous pulsars probably validate the
spherical wind model
The continuum parameters do not show any
significant variation over the orbit
Part II
U. Mukherjee & B. Paul 2005,
A&A, 431, 667
Be/X-ray Binaries
Courtesy : Negueruela
CircumstellatDisc
Formationnotclear
Probablyduetorapidrotation
Be stars : major subclass of the B stars
Be means a non-supergiant star of spectral class B
which shows emission lines
B stars : HeI absorption lines in optical spectrum
High Temp : He loses 1 e
He II Absorption lines : O stars : hottest of all stars
Accretion Barrier ?
Be phenomenon is episodic transient
No outburst NS in quiescence
Low accretion rate : r
m
> r
c
Centrifugal Barrier exists
X-ray transients like 4U 0115+63 (Campana et al. 2001with
BeppoSAX) &
GRO J1744-28 (Wijnands et al. 2002 with Chandra) in quiescence
with luminosities of ~10
33
erg s
-1
Also been thought to be in the centrifugally inhibited regime
3A 0535+262 : Accretion and pulsations
during quiescence
The transient pulsar :
Magnetic Field : 10
13
G
Pulse Period : 104 s
Orbital Period : 111 days
Optical Companion : O9.7
IIIe Be star
Orbital Geometry
3 obs with the Beppo-SAX on
2000 Sep 4 (5:14 UT) : A
Oct 5 (00:42 UT) : B &
2001 Mar 5 (22:52 UT) : C
Distance of 2 kpc for this system (Steele et al. 1998)
The 2--10 keV X-ray luminosities measured from the three
observations are in the range of (1.5-4.0) 10
33
erg s
-1
,
We detected pulsations in C down to 2 10
33
erg s
-1
At this accretion rate the system is expected to be in the
centrifugally inhibited regime
Useful on-source times for A, B & C : 32 ks, 39 ks & 51 ks
each for MECS and 19 ks, 30 ks & 43 ks for LECS
Data from only two of the MECS detectors were available
Timing Analysis
MECS light curves with res : 1.0 s
Average Count Rate for A & B
~ 0.04 count s
-1

For C ~0.09 count s
-1
Only C showed coherent pulsations
103.41 s
Another period : 101.57 s when
period search over a wider range
between (101--105) s
Long term pulse period history rules
out 101.57 s
Probably arose due to the window
function of the light curve
Extracted 5 separate light-curves each
with data only with avg count rate less
than 0.15 , 0.125 , 0.1 count s
-1
, 0.08 &
0.06 count s
-1
respectively
Detection of pulsations down to an
average count rate of less than
0.08 count s
-1
But pulse profile for 0.06 count s
-1
shows
pulsations (last panel)
Pulse profiles are background subtracted
Pulse Fraction = [(Maximum - Minimum)/
Maximum] ~ 50%
Thus it is fair to conclude that pulsations
exist for C down to at least a count rate of
0.06 count s
-1
Spectral Analysis
Spectra fitted with PL or BB + exp. line of sight absp
BB does not fit A, B or C : red
2
> 2.0
PL + exp. line of sight absp fits A, B well but not C
PL + BB + exp. line of sight absp fits C well
BB temp : 1.33 keV
Radius of BB ~0.1 km
The BB component !!
Quite different from the soft excess usually observed in some binary X-ray
pulsars (Yokogawa et al. 2000, Paul et al. 2002)
In general modeled as a BB with temp ~ 0.1 keV and the area of the
emitting region ~ 10
15 -16
sq. cm.
Hickox, Narayan and Kallman (2004) have explored
the physical origin of this so called soft excess
For sources with luminosity < 10
36
erg s
-1
, it may be due to emission by
photoionized or collisionally heated diffuse gas, or thermal emission from
the surface of the NS
This soft BB comp resembles that of 4U 0352+309 (Coburn et al. 2001)
& RX J0146.9+6121by Palombara & Mereghetti (2006)
This BB comp. has been attributed to polar cap emission
For all the 3 obs, A, B & C : accretion is expected to be
centrifugally inhibited
Detection of pulsations in C at these low flux levels indicates
that some matter may have leaked through onto the NS
surface
When the emission is non-pulsating, the X-ray spectrum is
PL type
Origin of an additional BB component in the pulsed
observation is not very clear though it seems to be essential
in fitting the spectra
Conclusions
Part III
U. Mukherjee & B. Paul, 2006c
U. Mukherjee & B. Paul, 2006,
J. Astrophys. Astr., 27, 37
B. Paul, H. Raichur & U. Mukherjee,
2005, A&A, 442, L15
Centaurus X-3 : Any reason behind the high-low
states ?
Detailed spectral analysis of the out of eclipse obs
with ASCA,BeppoSAX,Chandra,XMMNewton&RXTE
in its different intensity states
High & low states have separate domains of I
From RXTE : N
H
has a ceiling in the high states
Low states have high EQW
Compare our results with that of LMC X-4 & Her X-1
Properties of Cen X-3
P
s
~4.8 s
P
orb
~ 2.1 d
D ~ 8 kpc (Krzeminski 1974).
O-type supergiant V779 Cen Companion (Krzeminski 1974).
V779 Cen radius of ~12 R
S
& M ~17-19 M
S
(Hutchings et al.
1979).
QPOs at ~40 mHz (Takeshima et al. 1991; Audley et al. 1996)
strengthen the case for the presence of an accretion disk
Broadband (0.12-100) keV out-of-eclipse pulse-phase-
avg spec generally described by :
an abs PL + a broad Fe emission line @ ~ 6.7 keV +
cut-off @ ~ 14 keV (Santangelo et al. 1998, Burderi et
al. 2000)
Soft excess < 1 keV : BB with kT ~ 0.1 keV
(Burderi et al. 2000)
Cyclotron resonant feature @ ~ 28 keV
B ~ (2.4-3.0) 10
12
G (Santangelo et al. 1998)
RXTE-ASM in different energy bands shows that Cen X-3 has a
flux ~40 times more in the bursting state as compared to the low
state (Paul, Raichur & Mukherjee 2005)
Also, the low and high states last between a few to upto ~110 d
without having any periodicity
1.5 3.0 keV
3.0 5.0 keV
5.0 12.0 keV
Not only Cen X-3, but several other X-ray
Pulsars show long term X-ray intensity
modulations
Her X-1 & LMC X-4 show periodic X-ray
intensity variations, modulated at their
superorbital periods
Cen X-3 does not possess any such periodic
behaviour at all
There are complex X-ray spectral changes
associated with the intensity modulations at the
superorbital periods of Her X-1 & LMC X-4
(Naik & Paul 2003, 2004)
Broad band X-ray spectrum of Cen X-3 is more or
less similar to these pulsars
It is pertinent to explore the spectral changes of
this pulsar corresponding to its different intensity
states
Observations
80 obs with RXTE from 1996 to 2000 covering the various
intensity states
2 obs. With BeppoSAX on 1999-06-24 (High) & 2000-06-06
(Low)
XMM-Newton observed Cen X-3 on 2002-08-01 in the Low state
The obs with ASCA was on 1993-06-24 (Low)
With Chandra on 2000-12-30 (High)
All the observations were chosen in the out-of-eclipse state
Observation Log
RXTE Analysis
Continuum fitted with line of sight absp. + PL + a cut-off
In addition, a cyclotron absorption feature was also added
Fe-emission lines @ 6.4 keV, 6.67 keV and 6.95 keV
were also included in the form of gaussian lines
The centre energies were fixed to the laboratory values and the
widths of the gaussian lines were fixed to that obtained with the
XMM-Newton PN spectra respectively
All the other parameters were allowed to vary
A systematic uncertainty of 1.0% was added to the spectral
channels

For flux < 6.0 10
-10
erg cm
-2
s
-1
, there
seems to be a correlation between the
Cont. Flux & I
The spectrum becomes harder at lower
flux levels
At a flux level > 2 10
-9
erg cm
-2
s
-1
,
0.8 < I < 1.3
Chandra (H)
SAX (H)
SAX (L)
ASCA (L)
XMM (L)
Naik & Paul (2003) have reported from
RXTE that LMC X-4 is described in the
High state with 0.7 < I <1.0 &
Low state with 0.5< I <0.7
Thus for both these pulsars, a separate
range of values of I may be designated
for the low and high flux states
respectively
Variation of N
H
shows a few instances of high
to moderate values at low flux levels
For higher flux levels (> 2 10
-9
erg cm
-2
s
-1
),
there seems to be a ceiling of 4 10
22
cm
-2

For both I&

, there appears to be a lot of
scatter in their values as obtained with the
PCA for intermediate flux levels
ASCA (L)
SAX (L)
XMM (L)
SAX (H)
Chandra (H)
Variation of N
H
does not quite pesent a very
conspicuous picture either
Her X-1 shows an increase in N
H
by almost 2
orders of magnitude in its low state compared
to its high state over its 35-d superorbital
period (Naik & Paul 2003)
On the other hand, LMC X-4 does not show any
change in N
H
along its superorbital period
(Lang et al. 1998)
Hence we find that Cen X-3 shows results
which are in between these two extremes
Fe-line fluxes against (7 - 25)
keV PCA Flux does not show
any systematic trend (see figure
right)
Large & considerable scatter at
all flux levels for all the 3 line
intensities
EQW of the Fe-lines show a
reasonable trend (bottom panel,
right)
For low fluxes, the values of
EQW are quite high (a few
hundred eVs, especially for
Fe-Ko)
Then decreasing to acquire a
more or less steady value of
~100 eV for higher fluxes
(> 6 10
-9
erg cm
-2
s
-1
)
ASCA
SAX (H)
SAX (L)
Chandra
XMM
LMC X-4
Her X-1
Naik & Paul, 2003
A similar variation of the EQW with continuum flux was seen in case of
LMC X-4 & Her X-1
All the 3 sources show very high values of the EQW in the low
states whereas during the high states, no significant change is seen
Unlike LMC X-4 & Her X-1, Cen X-3 does not show any
global dependence of the iron-line intensities with the
continuum flux
Low states : it may seem Fe-line intensity has some
positive correlation with the continuum intensity, but
High states : no such dependence
LMC X-4 & Her X-1 : a clear correlation between the
fluxes at high states, indicating the production of the
iron line near the continuum X-ray source
Picture for Cen X-3 is far more intricate & it
suggests that not all the Fe-line is produced close
to the pulsar
Our work attempts to compare the X-ray
observational results for Cen X-3 in both its high
and low states with that of Her X-1 & LMC X-4
It is more or less established that Her X-1 &
LMC X-4 have an warped accretion disk which is
believed to be accounting for their flux behaviour
However spectral results obtained by us are not
fully in corroboration with that measured for the
the other 2 pulsars
Inference
Hence, based on these results it seems difficult to
put forward the same reasoning for the flux
behaviour in case of Cen X-3 as for the other two
Obscuration in the low states is present for
Cen X-3 (cases of high N
H
)
But that is not the only criterion to unequivocally
attribute the presence of high/low states to the
warped accretion disk
Ogilvie and Dubus (2001) : LMC X-4 & Her X-1 are 2
possible candidates to show a super-periodicity while
Cen X-3 is unlikely to exhibit such
Numerical simulations of the evolution of the
precessing accretion disks in XBPs :
Iping & Petterson (1990) propose for Cen X-3
a precessing accretion disk could describe a part
of its aperiodic X-ray flux variability
Day & Stevens (1993), based on EXOSAT data :
the central source accretes matter from a disk
which is fed by an X-ray excited wind emanating
from the companion
Scope for further theoretical and observational
work in this direction remains open
Especially extensive orbital coverage at lower
energies with detectors of the genre of Chandra
& XMM-Newton as well as a detailed broad band
coverage
BeppoSAX
Model : PL + BB + a gaussian line
as additive comp. & an abs along
the line of sight as the multipl.
comp.
Both the high and low state spectra
were well fitted
A broad Fe-emission line @
~6.65 keV
Presumably the blend of the
He-like triplet of Fe XXV
EQW (High) : 877 eV
EQW (Low) : 116 eV
High
Low
Model as used by
Burderi et al. (2000)
Fixed both the 6.44
keV and 6.96 keV
lines along with their
widths as given in
Burderi et al. (2000)

Red.
2
: 1.47 for
267 d.o.f.
~2.0 keV Si XIV
emission line
ASCA
Low
Iaria et al. (2005) have
analyzed the Chandra-HEG
spectrum in the (6.0 - 7.6)
keV & resolved the Fe XXV
triplet
We used only the first order
grating spectra
The zero-order source position determined by examining the HEG/MEG
dispersion lines & the data read out streak of the zeroth-order image
We used the same model as of Iaria et al. (2005) but allowed the cont.
params to vary
We froze the centre energies of the triplet and their respective widths
though we kept the params of the Fe Ko line thawed
Chandra
High
XMM-Newton
The 100 s binned PN-light curve
showed a broad coverage of eclipse,
egress and out of eclipse for about 68 ks.
Avg PN count rate was ~15 counts s
-1
For our spectral analysis, we
selected out-of-eclipse data
only (34 ks onwards)
Out of eclipse
Si XIV
S XVI
Ar XVII
+
Ar XVIII
Fe XXVI
Ca XX
Continuum Parameters
(*ASCA (L) & **SAX (H) : from Burderi et al. (2000))
4U 0114+65 : a luminosity dependent
study
Spectral characteristics of the high and low states
of the pulsar 4U 0114+65 with ASCA
Examine the change in the parameters of the
spectral model
Compare these spectral characteristics to the
results with other satellites
And with other X-ray pulsars ..........

AboutthePulsar..........
B1Ia supergiant optical companion
At a distance ~7 kpc (Reig et al. 1996)
Crampton, Hutchings & Cowley (1985) reported an
Orb Period ~11.59 d (optical radial velocity measurements)
Their measurements were not able to distinguish between a
circular or elliptical orbit
Corbet et al. (1999) obtained X-ray intensity modulations at a
period ~11.63 d from the long term light curve of
RXTE-ASM
X-ray light curve : considerable variability,
flaring activity for a few hours (Apparao et al. 1991,
Finley et al. 1992)
Also short-term flickering for minutes (Koenigsberger
et al. 1983)
Presence of a ~2.8 hr periodicity observed in the X-ray
light curves from EXOSAT and ROSAT data (Finley et al.
1992)
Farrell et al. (2005) reported the detection of a
superorbital period of ~30.7 d in this pulsar from
RXTE-ASM
Observations
Carried out by ASCA on 1997-02-10
Orbital phase~0.19
Useful exposure time ~25 ks
Total time span between start & end was
~55 ks (about 5.5% of the orbital period)
The intensity at the peak of the flare is ~ 12-15 times than the
persistent low level emission
At times, the intensity almost drops down to zero (@ ~22 & ~32 ks)
High state : 0-18 ks & rest : low state
GIS : 0.7 10.0 keV
SIS : 0.5 9.0 keV
Bin : 100 s
High
Low
Most significant difference between the
high & low state spectra is a change in
the Fe-line flux by a factor similar to the
change in the continuum
N
H
corroborates well with that of RXTE&Ginga
But SAX measures a high N
H
Masetti et al. (2005) reported for (1.5-100.0) keV whereas,
we are able to go down to 0.5 keV
Fe emission line @ 6.4 keV was reported with SAX, RXTE,
EXOSAT & Ginga
But the line was better detected during the low states
We report the detection of the Fe-line in both the states
with appreciable EQWs (Table 2)
ASCA has the best spectral resolution in comparison,
so the detectability of the Fe line can be supposed to
be most reliable
EQW of the Fe-line is found to increase appreciably in the low
states of several other pulsars like Her X-1 (Naik & Paul 2003),
LMC X-4 (Naik & Paul 2004) & SMC X-1 (Vrtilek et al. 2005)
High & low states of these pulsars are ascribed to the super-orbital
period precessingwarpedinneraccretiondisk
For 4U 0114+65, an abrupt decrease in X-ray flux over a few
thousand secs cannot be assigned to a super-orbital period
Rather, change in lum during our obs may be ascribed to a local
change in the density of the stellar wind
Rise & fall times of 1-2 ks as can be seen in the light curve
An assumed orbital velocity of a few hundred km/s indicates
clumpiness of the stellar wind at a length scale of about 10
10-11
cm
Summarising...........

Future Work ?!???!!
BinaryMillisecondXrayPulsars
Techniques used for HMXBPs can be directly applied
e.g. compare the characteristics of 3A 0535+262, to the
observations of BMXPs in quiescence
Furthermore, observations in the optical and IR will be
useful

PulsePhaseResolvedSpectroscopyof
LMXBslike4U162667withChandra
DetectionofCentrifugallyInhibited
AccretionforotherBe/Xraybinarieswith
BeppoSAX/XMMNewton