X-ray properties of the microquasar GRS 1915+105 during a variability class transition

Abstract

We present a detailed X-ray study of the microquasar GRS 1915+105 during a variability class transition observed in 2000 June with the PPCs of the Indian X-ray Astronomy Experiment. We supplement this observation with data from the RXTE archives. The source made a transition from a steady low-hard state to a regular oscillatory behaviour in the light curve known as bursts or class `rho' (Belloni et al. 2000) between 2000 May 11 and 17 and reverted back to the low-hard state on 2000 June 27. A gradual change in the burst recurrence time from about 75 s to about 40 s was observed which then increased to about 120 s during the ~ 40 days of class `rho'. The regular bursts disappeared from the X-ray light curves and the class transition was observed to occur within 1.5 hours on 2000 June 27 with the PPCs. A correlation is found between the observed QPO frequency at 5-8 Hz in the quiescent phase and the average X-ray intensity of the source during the class `rho'. We notice a strong similarity between the properties of the source during the class `rho' and those during the oscillatory phase of the observations of class `alpha'. From the timing and spectral analysis, it is found that the observed properties of the source over tens of days during the class `rho' are identical to those over a time scale of a few hundreds of seconds in the class `alpha'. Examining the light curves from the beginning of the RXTE/PCA and RXTE/ASM observations, it is found that the change of state from radio-quiet low-hard state to high state occurs through the X-ray classes `rho' and `alpha' which appear together during the state transition. It is further inferred that the source switches from low-hard state to the class `rho' through the intermediate class `alpha'. Comment: 10 pages with 9 figures, LaTex. To be appeared in MNRAS

Full text

Page 1

arXiv:astro-ph/0110589v1 28 Oct 2001
Mon. Not. R. Astron. Soc. 000, 000–000 (0000)Printed 1 February 2008(MN LATEX style file v1.4)
X-ray properties of the microquasar GRS 1915+105 during
a variability class transition
S. Naik, P. C. Agrawal, A. R. Rao, and B. Paul
Tata Institute of Fundamental Research, Homi Bhabha Road, Mumbai, India 400 005
sachi@tifr.res.in (SN), pagrawal@tifr.res.in (PCA), arrao@tifr.res.in (ARR), bpaul@tifr.res.in (BP)
Accepted for publication in MNRAS, 2001
ABSTRACT
We present a detailed X-ray study of the microquasar GRS 1915+105 during a vari-
ability class transition observed in 2000 June with the Pointed Proportional Counters
(PPCs) of the Indian X-ray Astronomy Experiment (IXAE). We supplement this ob-
servation with data from the RXTE archives. The source made a transition from a
steady low-hard state to a regular oscillatory behaviour in the light curve known as
bursts or class ρ (Belloni et al. 2000) between 2000 May 11 and 17 and reverted back
to the low-hard state on 2000 June 27. A gradual change in the burst recurrence time
from about 75 s to about 40 s was observed which then increased to about 120 s during
the ∼ 40 days of class ρ. The regular bursts disappeared from the X-ray light curves
and the class transition was observed to occur within 1.5 hours on 2000 June 27 with
the PPCs. A correlation is found between the observed QPO frequency at 5−8 Hz in
the quiescent phase and the average X-ray intensity of the source during the class ρ.
We notice a strong similarity between the properties of the source during the class ρ
and those during the oscillatory phase of the observations of class α which contains
a long stretch (∼ 1000 s) of steady low-hard state in the light curve along with the
regular periodic bursts. From the timing and spectral analysis, it is found that the
observed properties of the source over tens of days during the class ρ are identical to
those over a time scale of a few hundreds of seconds in the class α. Examining the
light curves from the beginning of the RXTE/PCA and RXTE/ASM observations, it
is found that the change of state from radio-quiet low-hard state to high state occurs
through the X-ray classes ρ and α which appear together during the state transition. It
is further inferred that the source switches from low-hard state to the class ρ through
the intermediate class α.
Key words:
stars: individual: GRS 1915+105 — X-rays: stars
accretion, accretion discs — binaries: close — black hole physics —
1INTRODUCTION
The Galactic X-ray source GRS 1915+105, discovered in
1992, was identified with a superluminal radio source at a
distance of 12.5 ± 1.5 kpc (Mirabel & Rodriguez 1994). Its
radio characteristics like jets and superluminal motion are
similar to those found in quasars and hence this source is
called a microquasar. The source is very bright in X-rays
and shows strong variability over a wide range of time scales.
The X-ray emission is characterized by quasi-periodic oscil-
lations (QPOs) at centroid frequencies in the range of 0.001
− 67 Hz (Morgan, Remillard, & Greiner 1997). Based on ex-
tensive X-ray studies, Muno, Morgan, & Remillard (1999)
classified the behavior of the source into two distinct states,
a spectrally hard-state with the presence of narrow QPOs,
dominated by a power-law component and a soft-state with
the absence of QPOs, dominated by thermal emission.
Belloni et al. (2000) have made an extensive study of
the X-ray emission of the source and classified all the pub-
licly available RXTE/PCA observations from 1996 January
to 1997 December into 12 different classes on the basis of
structure of the X-ray light curve and the nature of the
color-color diagram. They found that the source variability
is restricted into three basic states, a low-hard state with in-
visible inner accretion disk (C), a high-soft state with visible
inner accretion disk (B) and a low-soft state with spectrum
similar to the high-soft state but with much less intensity
(A). However, in GRS 1915+105 the observed fast and slow
transitions from one state to other are not clearly under-
stood.
The different variability classes of Belloni et al. (2000)
c ? 0000 RAS

Page 2

2
Naik, S. et al.
range from steady emission for long durations like class φ
(state A), class χ (state C) to large amplitude variations
(from state C to state B with a hint of state A) in classes λ,
κ, ρ. In the high state (state B), the source sometimes shows
short periodic flickering with different amplitudes (classes γ,
µ and δ). During the classes θ, β and ν the amplitude varia-
tion is accompanied by soft X-ray dips (state A) with dura-
tion of a few tens of seconds to hundreds of seconds. It has
been suggested that these soft dips in X-ray light curves are
also associated with radio flares (Naik et al. 2001). The class
χ is further divided into four sub-classes χ1, χ2, χ3 and χ4,
depending on the count rate and hardness ratio in different
energy ranges. During the class ρ, the X-ray light curve con-
sists of regular and characteristic pattern known as bursts
(Taam, Chen, & Swank 1997; Yadav et al. 1999) which is re-
flected as a loop-like behaviour (Vilhu & Nevalainen 1998) in
the color-color diagram. Similar type of oscillations are ob-
served in the X-ray light curve for a duration of a few hun-
dreds of seconds followed by a long quiet period of about
1000 s as seen in the class χ2/χ4. The observations with
these characteristics are as classified into class α. Among
the twelve different classes of X-ray observations, it is seen
that the source is radio-loud in X-ray classes θ, β, χ1 and
χ3 (Naik & Rao 2000).
We present theresults
GRS 1915+105 with the PPCs onboard the IXAE when the
source made a class transition from a regular oscillatory be-
haviour of class ρ to a steady state of class χ2/χ4. We have
also analyzed the available RXTE archival data contempo-
raneous to the PPC observation. In the following sections
we present results of these studies.
oftheobservationof
2INSTRUMENT AND OBSERVATIONS
The X-ray observations of the microquasar GRS 1915+105
were made using the PPCs of the IXAE on board the In-
dian satellite IRS-P3. The IXAE includes three co-aligned,
identical, multi-wire, multi-layer proportional counters with
an effective collecting area of 1200 cm2and a field of view
of 2.3◦× 2.3◦. All the PPCs operate in the energy range
of 2−18 keV with an average detection efficiency of about
60% in 3−7 keV energy range. For a detailed description
of the PPCs refer to Agrawal (1998) and Rao et al. (1998).
Background observations were made before and after the
source observation by pointing the PPCs to a source-free
region in the sky, close to the target source. The source
GRS 1915+105 was observed from 2000 June 18 to 22 with
1 s time integration mode and from June 23 to 27 in 0.1 s
integration mode for a total useful period of 29460 s. The
log of the PPC observations is given in Table 1.
Regularpointedobservations
GRS 1915+105 during 1996 − 2001 period with the Rossi X-
ray Timing Explorer (RXTE) satellite provide a good cover-
age for a detailed study of the timing and spectral properties
of the source and establish the associated radio properties
with X-ray emission during different spectral states. There
have been over 550 observations of GRS 1915+105 with the
PCA and HEXTE of the RXTE. The list of observations
used in the present work is given in Table 2.
ofthe microquasar
Figure 1. The X-ray light curves for GRS 1915+105 obtained
with the PPCs (PPC−1 and PPC−3 added) in the energy range
2−18 keV in 1 s time integration mode from the observations in
2000 June. The date and orbit of the observations are indicated
in each panel of the figure.
3ANALYSIS AND RESULTS
3.1Data from IXAE Observations
The X-ray data in the energy band 2−18 keV and 2−6 keV
were corrected for pointing offset using the aspect informa-
tion. The background count rates obtained by pointing the
detectors to a source-free region, were subtracted from the
source count rate. Dead time correction has been neglected
as it is less than 1% even at a count rate of about 500 counts
s−1per PPC. From calibration of the PPCs using Crab Neb-
ula, it is found that the spectral data from PPC−3 is more
reliable. Hence the hardness ratio (ratio of the count rates
in 6−18 keV and 2−6 keV bands) has been obtained using
data only from PPC−3.
We have generated the X-ray light curves of the source
in 2−18 keV range for 500 s using the data from PPC−1
and PPC−3 for each orbit of observation. The light curves
for a few orbits are shown in Figure 1. The light curves show
that low intensity quiescent phase lasting for a few tens of
seconds, is followed by the bursts, which have exponential
increase in X-ray flux with rise time of ∼ 8−10 s, high peak
flux level for ∼ 10 s and a sharp linear decay in ∼ 2−3
s. The average burst recurrence time (the time taken by a
c ? 0000 RAS, MNRAS 000, 000–000

Page 3

X-ray properties of the microquasar GRS 1915+105 during a variability class transition
3
Figure 2. The X-ray light curves of GRS 1915+105 obtained from PPC−3 of IXAE during four different intervals of 2000 June
observation are shown in the energy ranges 2−18 keV and 2−6 keV along with hardness ratio H. R. (ratio of count rate in 6 − 18 keV
range to count rate in 2−6 keV range). The appearance of secondary (and tertiary peaks) can be seen in the later part of the observation.
cycle of quiescent interval followed by a burst) for all the
orbits of the PPC observations and the burst strength (the
ratio between the peak count rate of the burst and the qui-
escent count rate at 2−18 keV for PPC−1 and PPC−3) are
given in Table 1. An increase in the burst recurrence time
by a factor of about 2 from the beginning to the end of the
observations is clearly detectable from the table. The X-ray
light curves in 2−18 keV and 2−6 keV bands along with the
hardness ratio obtained from PPC−3 data for four different
PPC observations are shown in Figure 2. It can be seen that
the spectrum becomes hard as the burst progresses and it is
hardest at the end of the decay of the burst. Presence of sec-
ondary (and tertiary) peaks during the decay phase of the
bursts in the X-ray light curves is also discernible. From the
light curves and hardness ratio for the third orbit of PPC
observations on June 27 shown in the fourth panel of Figure
2, it can be seen that the regular bursts disappeared in both
the energy bands and the transition from class ρ to class
χ2/χ4 occurred within about 1.5 hours, the orbital period
of IRS-P3 satellite.
The burst strength has been calculated from a typical
burst from each PPC observation. The peak count rate was
determined as the average count rate over a time span of
4 seconds during the peak of the bursts and the quiescent
count rate as the average count rate over 20 seconds just
after the decay of a burst. The variation of the burst re-
currence time with the burst strength is shown in Figure
3. An anti correlation is found between the count rate ra-
tio (burst strength) and the average burst recurrence time
with a correlation coefficient of -0.54 (30 degrees of free-
dom). This indicates that the bursts were stronger during
the early phase of observations when the burst recurrence
time was small and the burst strength decreased with the
increase in burst recurrence time.
3.2 Data from RXTE Observations
To corroborate the IXAE results we have also analyzed 31
RXTE/PCA observations of GRS 1915+105 made between
2000 May 4 and 2000 July 5 which are listed in Table 2.
Other details like the average count rate during quiescent
phase and burst phase, burst recurrence time, the observed
low frequency QPOs (∼ 2−8 Hz), hardness ratios and rms
variability are also given in the table. The hardness ratio
HR1 is described as the ratio of the count rates in the en-
ergy range 5−13 keV to that in 2−5 keV whereas the ratio
c ? 0000 RAS, MNRAS 000, 000–000

Page 4

4
Naik, S. et al.
Table 1. Log of X-ray Observation of GRS 1915+105 with the
PPCs of IXAE
Observation
Date1
Start time
(UT)
End Time
(UT)
BRT2
(s)
Burst
strength3
orbit
181
4
1
2
3
4
5
1
2
3
4
5
2
3
4
5
4
5
2
3
1
2
3
1
2
3
1
2
1
2
3
13:00
18:04
12:44
14:20
16:02
17:43
19:35
12:16
14:00
15:41
17:23
19:12
13:42
15:21
17:03
18:49
16:41
18:27
16:21
18:00
14:25
15:59
17:39
14:07
15:37
17:17
15:17
16:58
15:01
16:36
18:18
13:18
18:23
12:54
14:38
16:20
18:02
19:43
12:36
14:16
16:00
17:42
19:22
13:57
15:38
17:21
19:02
16:59
18:40
16:37
18:19
14:34
16:16
17:58
14:13
15:54
17:36
15:34
17:16
15:12
16:54
18:35
53
48
46
46
50
51
47
47
49
48
46
49
43
48
53
52
54
57
49
49
77
75
93
85
107
92
80
98
54
126
− − −−
2.45±0.15
2.87±0.16
2.95±0.16
2.79±0.14
3.14±0.17
2.87±0.21
2.93±0.15
3.24±0.17
2.81±0.16
3.42±0.20
3.09±0.19
3.40±0.17
3.19±0.18
3.55±0.19
3.57±0.22
3.33±0.18
3.30±0.18
2.55±0.16
2.80±0.18
2.94±0.28
3.60±0.19
2.89±0.14
2.61±0.13
3.02±0.13
2.75±0.12
2.49±0.11
2.67±0.12
2.51±0.12
1.77±0.08
2.34±0.12
− − −−
19
20
21
22
23
24
25
26
27
1Date − 2000 June,
2BRT − Burst Recurrence Time (averaged over the orbit)
3Burst strength − Ratio of Peak count rate to quiescent count
rate
of the count rates in the energy bands of 13−60 keV and
2−13 keV is described as HR2. The source was in a low-
hard state (χ2/χ4) on and before 2000 May 11 and changed
to the oscillatory nature of class ρ and remained for ≥ 40
days (from 2000 May 17 to 2000 June 27, as observed with
RXTE/PCA). It returned back to the low-hard state which
was detected on 2000 July 5. The burst recurrence time for
each of the RXTE observation of class ρ was calculated us-
ing the PCA X-ray light curves. A plot of the average burst
recurrence time versus the observation date (MJD) for the
combined IXAE and RXTE/PCA observations is shown in
Figure 4. It can be seen that the burst recurrence time, ob-
tained from the RXTE was about 75 s in the beginning and
it gradually decreased to ∼ 40 s. Towards the end of the ob-
servations, regular monitoring by PPCs revealed a gradual
increase in the burst recurrence time to ∼ 120 s which is
followed by the disappearance of the bursts from the light
curve. This indicates the switching of the source from oscil-
latory behaviour of class ρ to a radio-quiet steady emission
of class χ2/χ4.
Figure 3. Plot of the average burst recurrence times obtained
from IXAE versus the ratio of the peak flux and quiescent flux in
2 − 18 keV energy is shown. A negative correlation between the
burst duration and the ratio between the peak flux and quiescent
flux is seen.
We have used the 8 ms time resolution PCA data in
2−13 keV band to generate the power density spectra (PDS)
for all the RXTE/PCA observations listed in Table 2. Dur-
ing the low-hard X-ray states (on 2000 May 4 and 2000 July
5, class χ2/χ4), QPOs are detected at frequencies 2.8−3.9
Hz whereas during the rest of the PCA observations (class
ρ) QPOs are observed at higher frequencies (6−8 Hz). A
plot of the QPO frequency and the source flux in 2−60 keV
band for 2 PCUs during the observations of class ρ, is shown
in Figure 5. A strong correlation is found between the QPO
frequency and the average X-ray flux with a correlation co-
efficient of 0.85 (for 25 degrees of freedom).
Various X-ray characteristics of GRS 1915+105 such
as nature of the light curve, gradual change in the burst
recurrence time, hardness ratios HR1 and HR2, presence
of QPOs and spectral properties in class ρ are found to be
similar to those in the oscillatory phase of class α (Belloni et
al. 2000). These properties are described in the next section.
4COMPARISON BETWEEN CLASSES
ρ AND α
4.1Timing Analysis
To compare the X-ray timing properties of GRS 1915+105 in
the X-ray classes ρ and α, we have selected one RXTE/PCA
observation of each class (Obs. IDs 20402-01-03-00 and
20402-01-28-00 respectively) when all the 5 PCUs were on.
Figure 6 shows the X-ray light curve for the above two
RXTE/PCA observations for 1 s time bin obtained from the
standard-1 data. The upper panel shows the presence of reg-
ular bursts with recurrence time of ∼ 60 s in the X-ray light
curve of class ρ whereas bursts with increasing recurrence
time and decreasing peak intensity followed by a low-hard
state lasting for ∼ 1000 s (class α) are shown in the bottom
panel of Figure 6. At the end of the long stretched (∼ 1000
s) low-hard state,the bursts again reappear with a very high
c ? 0000 RAS, MNRAS 000, 000–000

Page 5

X-ray properties of the microquasar GRS 1915+105 during a variability class transition
5
Table 2. List of RXTE/PCA observations analyzed from the archival data
RXTE Observation1
Average Countrate (2 PCUs)
QP2
BP3
QPO frequency
(in Hz)
BRT4
(in s)
Quiescent phase
HR15
HR26
IDDateTotalRMS
class ρ
50703-01-10-00
50703-01-10-01
50703-01-10-02
50703-01-11-00
50703-01-11-01
50703-01-11-03
50703-01-12-00
50703-01-12-01
50703-01-12-02
50703-01-12-03
50703-01-13-00
50703-01-13-01
50703-01-13-02
50703-01-13-03
50703-01-14-00
50703-01-14-01
50703-01-14-02
50703-01-15-00
50703-01-15-01
50703-01-15-02
50703-01-15-03
50703-01-16-01
50703-01-16-00
50703-01-16-02
50703-01-16-03
20402-01-03-00♣
05/17/2000
05/17/2000
05/17/2000
05/25/2000
05/25/2000
05/25/2000
05/31/2000
05/31/2000
05/31/2000
05/31/2000
06/07/2000
06/08/2000
06/08/2000
06/08/2000
06/14/2000
06/15/2000
06/15/2000
06/19/2000
06/20/2000
06/20/2000
06/20/2000
06/26/2000
06/26/2000
06/27/2000
06/27/2000
11/19/1996
3317
3182
3088
2894
3054
2724
2642
2827
2672
2540
2859
2908
2652
2668
3875
3617
3521
2777
3066
2967
2968
5226
5483
4941
5274
4987
10684
9651
9207
9514
9264
8456
9443
10104
10554
9626
9415
9838
8601
9668
9182
9523
9004
8794
10068
10463
9910
11574
10697
10639
9926
9863
3822
3703
3560
3466
3556
3344
3320
3365
3340
3183
3517
4703
3467
3495
5035
4237
4140
5206
5301
5399
5235
6217
6332
5455
5717
5598
6.28±0.062
6.67±0.079
6.48±0.076
6.62±0.138
6.44±0.077
6.15±0.087
6.41±0.067
6.42±0.053
6.32±0.064
6.48±0.146
6.74±0.0515
6.60±0.0435
6.56±0.0585
6.68±0.085
6.74±0.0295
6.83±0.0435
6.80±0.0335
6.81±0.045
6.70±0.0535
6.91±0.06
6.84±0.043
7.22±0.0305
7.34±0.0395
7.27±0.082
7.23±0.05
7.46±0.011
73.68
66.67
66.67
55.56
56.00
50.00
46.67
50.00
47.37
46.15
42.50
46.67
47.06
46.15
57.14
54.55
53.85
46.51
49.18
46.51
48.72
80.65
70.59
72.73
73.33
76.92
0.68
0.66
0.66
0.58
0.57
0.57
0.51
0.52
0.52
0.52
0.48
0.50
0.49
0.49
0.61
0.59
0.59
0.48
0.48
0.47
0.47
0.57
0.58
0.57
0.57
1.03
0.055
0.053
0.059
0.056
0.059
0.058
0.059
0.059
0.056
0.056
0.053
0.054
0.051
0.049
0.051
0.051
0.051
0.048
0.048
0.046
0.049
0.046
0.043
0.047
0.051
0.068
0.068
0.062
0.064
0.068
0.065
0.064
0.056
0.067
0.062
0.063
0.047
0.057
0.063
0.066
0.066
0.061
0.052
0.056
0.049
0.052
0.054
0.065
0.062
0.069
0.065
0.067
class α and χ2
20402-01-28-00♣
20402-01-04-00♣
50703-01-08-00
50703-01-08-01
50703-01-09-00
50703-01-09-01
50703-01-17-00
50703-01-17-01
05/18/1997 (α)
11/28/1996 (χ2)
05/04/2000 (χ2)
05/04/2000 (χ2)
05/11/2000 (χ2)
05/11/2000 (χ2)
07/05/2000 (χ2)
07/05/2000 (χ2)
3442
—–
—–
—–
—–
—–
—–
—–
5480
—–
—–
—–
—–
—–
—–
—–
——
4414
3119
3226
2977
2703
3538
3302
————
4.57±0.016
3.37±0.016
3.61±0.017
3.96±0.025
3.82±0.0305
2.97±0.0125
2.78±0.0125
——–
——–
——–
——–
——–
——–
——–
——–
1.07
1.16
0.87
0.86
0.775
0.78
0.752
0.766
0.085
0.099
0.084
0.082
0.078
0.082
0.077
0.08
0.084
0.116
0.141
0.137
0.124
0.133
0.164
0.168
1Data used for spectral analysis are indicated by ♣
2QP : Quiescent phase,3BP : Burst phase
4BRT: Burst recurrence time (averaged over the orbit of observation)
5HR1: Ratio of count rate in 5 − 13 keV to count rate in 2 − 5 keV
6HR2: Ratio of count rate in 13 − 60 keV to count rate in 2 − 13 keV
peak X-ray flux (similar to the peak flux of class ρ) which
decreases gradually with increasing burst duration. We have
shown, in Figure 7, the X-ray light curve for 1 s time bin in
2−60 keV energy range along with the hardness ratios (HR1
and HR2) for classes ρ (left panels) and α (right panels). It is
seen that all the properties (structure of the light curves and
the hardness ratios) are identical during the burst/quiescent
phase of the two X-ray classes. The presence of QPOs in the
PDS during the burst/quiescent phase of the classes ρ and α
has been shown by Muno et al. (1999) (Fig. 1 (d) and 1(j)).
From those figures, it can be seen that the QPO frequency
during the particular observation of class ρ lies between 5−7
Hz. During the class α, the QPO frequency changes from ∼
8 Hz at the beginning of the burst phase, to ∼ 5 Hz during
the change of state to the low-hard state. As the QPO fre-
quency varies directly with the intensity, the change in QPO
frequency during the burst phase of class α is due to a de-
crease in the X-ray intensity with the increase of recurrence
time. In both the classes, the source is radio-quiet with sim-
ilar flux densities at 2.25 GHz and 8.3 GHz (Naik and Rao
2000).
From the IXAE and RXTE observations of the source
during 2000 May−June (when the source was in the X-ray
class ρ), it is noticed that there is a gradual change in the
various properties of the source like burst recurrence time,
intensity, QPO frequency etc. and the source makes a grad-
ual transition to the low-hard state of the class χ2/χ4. These
characteristics are very similar to those observed in the ob-
c ? 0000 RAS, MNRAS 000, 000–000

Page 6

6
Naik, S. et al.
Figure 4. The average burst recurrence times obtained from
IXAE and RXTE/PCA observations during the long stretch of
X-ray class ρ in 2000 May − June are plotted with the day of ob-
servation of the source (MJD). A change in burst duration with
observation dates can be clearly seen.
servations of class α but at a much shorter time scale. Hence,
we argue that the nature of the source over time scale of a
few tens of days during the class ρ is identical to that over
a time scale of ∼ 1000 s during the class α.
4.2Spectral Analysis
To compare the results with earlier observations of the
classes ρ, α and χ2/χ4, we have selected RXTE pointed
observations of each class when all 5 PCUs were on. The
observations made on 1996 November 19 (20402-01-03-00)
and 1997 May 18 (20402-01-28-00) are used for classes ρ
and α respectively whereas the observation of 1996 Novem-
ber 28 (20402-01-04-00) is used for the class χ2. Details of
these observations are given in Table 2. To study the spec-
tral behaviour of the source during different phases, we have
selected data at suitable time ranges for the burst/quiescent
phases of the source in classes ρ and α. For the class ρ, we
have selected data for the burst phase when the source count
rate was ≥ 15000 counts s−1and for the quiescent phase
when the source count rate was ≤ 10000 counts s−1. For
the class α, the burst phase corresponds to the source count
rate ≥ 12000 and the quiescent phase to a count rate in the
range of 7500 to 9500 (for 5 PCUs). To compare the spec-
trum of the source during the long stretch (∼ 1000 s) of low-
hard state in class α with that of class χ2, we have selected
the data when the source count rate was ≤ 6000. To get a
complete spectrum in the broad energy band (3−150 keV),
we have combined the data obtained from RXTE/PCA and
HEXTE. Energy spectra in 129 channel were generated from
the Standard 2 mode PCA data. Standard procedures were
applied for the data selection, background estimation and
response matrix generation (Rao et al. 2000). Systematic
error of 2% have been added to the PCA spectral data. For
Figure 5. The variation of QPO frequency with the average X-
ray count rate (for 2 PCUs) during the class ρ in 2000 May−June
from RXTE/PCA observations is shown. A direct correlation be-
tween the average QPO frequency with the average X-ray lumi-
nosity during the class ρ is apparent in the figure.
Figure 6. The light curve of GRS 1915+105 obtained from the
RXTE/PCA data on MJD 50406 (class ρ) in 2 − 60 keV energy
range for 1 s time bin is shown in the upper panel. The lower
panel shows the light curve of the source on MJD 50586 (class α)
in the same energy range and time bin. The presence of regular
bursts is clearly seen in the upper panel and bursts with increasing
duration are seen in the lower panel.
HEXTE, standard 2 as well as the archive mode data from
the HEXTE Cluster 0, which has better spectral response,
have been used.
We have fitted the spectrum for the burst and the qui-
escent phases of both the classes ρ and α using the stan-
dard black hole model (Muno et al. 1999) consisting of
“disk-blackbody and power-law” and “disk-blackbody and
a thermal-Compton spectrum” with a fixed value of absorp-
c ? 0000 RAS, MNRAS 000, 000–000

Page 7

X-ray properties of the microquasar GRS 1915+105 during a variability class transition
7
Figure 7. The X-ray light curves of GRS 1915+105 in the 2 − 60 keV range for classes ρ and α are shown (panels a and d respectively)
along with plots of hardness ratios HR1 (count rate in 5 − 13 keV / count rate in 2 − 5 keV energy range) in panels (b) and (e) and
HR2 (count rate in 13 − 60 keV / count rate in 2 − 13 keV energy range) vs time in panels (c) and (f) respectively. From the figure, it
is seen that the burst properties in class α are similar to those seen in class ρ.
tion by intervening cold material parameterized as equiva-
lent Hydrogen column density NH at 6 × 1022cm−2. Simul-
taneous spectral fits to the PCA (in the energy range of 3−25
keV) and HEXTE (in the energy range of 15−150 keV) spec-
tra were performed for different phases of classes ρ and α,
keeping the relative normalization as a free parameter. The
same procedure was followed for the fitting of the spectrum
during the long stretched low-hard state of class α and the
observation of class χ2. The fitted parameters for the above
two models are given in Table 3. From the table, it is ob-
served that for the “disk-blackbody and power-law” model,
there is no significant difference in the spectral parameters
for the two different classes. For the “disk-blackbody and
thermal-Compton spectrum” model, it is seen that the disk
temperatures, and the optical depth of the Compton cloud
τ are identical for the classes α and ρ during the differ-
ent states. Comparing these two models, it is found that the
model with disk-blackbody and thermal-Compton spectrum
fits better than the model with disk-blackbody and power-
law.
The resultant deconvolved energy spectra for the “disk
blackbody and thermal-Compton” model are shown in Fig-
ure 8 for different phases of the classes α and ρ and the
long stretched low-hard states of classes α and χ2. From the
figure and Table 3, it is clear that the source spectrum is
identical during the burst and quiescent phases of the two
X-ray classes α and ρ.
4.3ASM Light Curve
We have examined the X-ray light curves of 562 publicly
available RXTE/PCA observations till 2000 October 11. We
found that 70 observations are of class ρ and 30 observations
are of class α. Naik and Rao (2000) have described the sim-
ilarities in the radio properties of the source during both
the ρ and α classes. To examine the pattern of change of
state of the source between the X-ray classes ρ and α, we
show in Figure 9 the one day averaged ASM light curve of
GRS 1915+105 with the identified ρ and α classes marked
with filled circles and filled triangles respectively. From the
figure, it is noticed that the change of state of the source
occurs from the radio-quiet low-hard state (χ2/χ4) to high-
state through the classes α and ρ. It is also observed that
these two classes appear together during the switching of
state from low-hard state to high state which strengthen
the argument that the classes α and ρ have strong similar-
ity. During the switching of the source from a low-hard state
(χ2/χ4) to the high-state, it goes through the class α fol-
lowed by class ρ which is very clear during the transition
from a long stretch of low hard state of class χ2/χ4 to high
state during MJD 50490 − 50600. Direct switching to the
class ρ from the low-hard state is not observed.
Based on the observed X-ray and radio properties, we
find that the X-ray classes ρ and α are identical except for
the presence of a long stretch of low-hard state in class α
similar to the radio-quiet low-hard states of class χ2/χ4. In
other-words, we can say that the X-ray class α is the same
as the combined class of ρ and χ2/χ4.
5DISCUSSION
The Galactic microquasar GRS 1915+105 shows extended
low-hard states on several occasions. These low-hard states
are characterized by a low frequency QPO at ∼ 3 Hz. The
c ? 0000 RAS, MNRAS 000, 000–000

Page 8

8
Naik, S. et al.
Figure 8. The deconvolved X-ray spectra of GRS 1915+105 during the burst and quiescent phases of the classes α and ρ and the
low-hard states of classes α and χ2. The fitted model consists of a disk blackbody and thermal Compton components.
source switches from an extended low-hard state into a high-
soft state in a wide range of time scales. Rao et al. (2000)
have observed a slow transition from a low-hard state to a
high-soft state (in about 3 months) in 1997 March−August.
They have also reported a state transition in a very fast
timescale (a few seconds) when the source was exhibiting
irregular bursts in the X-ray light curves. However, the ob-
served state transition of the source over a wide range of
time interval is not yet clearly understood. On 2000 June
27, the source made a transition from a regular oscillatory
behaviour (class ρ, in which the light curve contains reg-
ular bursts with recurrence time in the range 40−120 s)
to a steady emission (low-hard state) within 1.5 hour. The
characteristic properties of slow rise and fast decay of these
regular bursts of class ρ are unique. Taam et al. (1997) have
attempted to describe these bursts in the framework of ther-
mal/viscous instabilities in the accretion disk. Vilhu and
Nevalainen (1998) tried to explain the properties of these
regular bursts by using a two phase self-consistent thermal
radiative model. Paul et al. (1998) interpreted the observed
slow rise and fast decay of the regular bursts as the evidence
for the disappearance of matter into the event horizon of
the black hole. Chakrabarti et al. (2000) interpreted the ob-
served features in GRS 1915+105 in the light of advective
disk paradigm which includes self-consistent formulation of
shocks and out-flows from post-shock region and described
the oscillatory behaviour during the X-ray classes ρ and α
as the combination of the high count rate “On-state” and
the low-count rate “Off-state”. Using this model, we have
tried to explain the observed correlation between the QPO
frequency and the X-ray intensity during the observations
of class ρ.
Figure 9. The ASM light curve of GRS 1915+105 obtained from
RXTE/ASM from one day averaged dwell data. The presence of
X-ray classes ρ and α are indicated in the figure by filled circles
and filled triangles respectively.
Chakrabarti & Manickam (2000) derived a relation be-
tween the QPO frequency in 1−10 Hz range and duration of
the quiescent phases (class ρ and α). As the 1−10 Hz QPO
could be due to the oscillation of shocks located at tens to
hundreds of Schwarzschild radius, they tried to explain the
switching of burst phase to quiescent phase and vice versa
c ? 0000 RAS, MNRAS 000, 000–000

Page 9

X-ray properties of the microquasar GRS 1915+105 during a variability class transition
9
Table 3. Spectral parameters during classes α, ρ and χ2/χ4
Parameters1
Quiescent phaseBurst phaseLow-hard state
αραραχ2
Model: Disk blackbody (diskbb) + power-law
Reduced χ2
Γx
kTin(keV)
HEXTE Count rate
PCA Count rate
1.49
2.71±0.02
1.31±0.02
63±1
8001±37
1.97
2.86±0.02
1.44±0.02
63±1
9474±45
1.49
2.75± 0.03
1.8 ± 0.01
68±1
12647±59
1.34
2.80 ± 0.01
1.99 ± 0.01
62±1
13994±64
1.87
2.39±0.02
1.05±0.04
61±1
4336±19
2.79
2.76±0.01
1.31±0.2
93±1
9973±44
Model: Disk blackbody (diskbb) + thermal-Compton (CompST)
Reduced χ2
kTin(keV)
kTe (keV)
τ
0.99
1.32 ± 0.02
19.45 ± 2.1
2.87 ± 0.23
1.53
1.37 ± 0.02
13.14 ± 1.28
3.49 ± 0.26
1.46
1.79 ± 0.02
57.45 ± 28.4
1.24 ± 0.49
1.30
1.96 ± 0.01
30.39 ± 7.85
1.96 ± 0.37
0.79
1.17 ± 0.03
15.73 ± 1.08
3.91 ± 0.2
0.9
1.35 ± 0.03
21.82 ± 1.2
2.66 ± 0.1
1: Γx : Power-law photon index, kTin: Inner disk temperature, kTe : Temperature of the Compton cloud, τ : Optical depth of the
Compton.
and the duration of the quiescent phase by assuming an
outflow from the post-shock region. Assuming the flow to
be isothermal until Rc (the location of the sonic point), a
shock compressed gas with R (the shock compression ratio)
> 1, must produce outflows which pass through the sonic
points. Hence matter is subsonic until the sonic point. As the
Compton cooling becomes catastrophic when the Thomson
scattering opacity (kes) becomes 0.4, the duration of the off
state (i.e. the duration between the end of a burst and the
beginning of the next burst which we described as the qui-
escent phase) is given by toff = (4πR2
the relation between inflow and outflow rates (Chakrabarti
1998), α = 3/2 (for a low angular momentum freely falling
matter), and v0 = 1, the duration of the off state is given
by
c)/(3 ˙ Mout kes). Using
toff = 461.5 (0.1
Θ ˙
M
) (
M
10M⊙)−1/3ν−4/3s
(1)
where
Θ ˙
M=Θout
Θin
˙ M
˙ MEdd
(2)
Θin and Θout are the solid angles of the inflow and the out-
flow respectively. Keeping the numerical coefficient constant
in the above equation by putting the value of α = 1 and v0
= 0.066, which is very reasonable for a black hole accretion,
one obtains
toff = 461.5 (0.1
Θ ˙
M
) (
M
10M⊙)−1(
v0
0.066)2ν−2s
(3)
This relation has been found to be valid for all the ob-
servations. According to this relation, the greater is the value
of the duration of the off state toff, less is the frequency of
the QPOs. If one compares the presence of QPOs and the
duration of the off state (quiescent phase) in the Figure 1(d)
of Muno et al. (1999), it can be seen that, during class ρ the
QPOs are at similar frequencies. Also in their Figure 1(j),
during the on and off states of class α, the QPO frequencies
change according to the above relation.
In class ρ, the relation between QPO frequency and
toff should be valid for a given day and the proportionality
constant varies for each day. We find that toffν2is correlated
with the quiescent count rate with a correlation coefficient
of 0.94. This can be explained by the fact that for a given
day Θ ˙
the quiescent count rate.
Althoughdifferentstates
GRS 1915+105 are explained by various models, none of
the models explain the transition from one state to other.
It would be interesting to study the changes in the emission
processes in the accretion disk and the factors which trigger
the change of state as well as the transition from one class
to another in the source.
Mcan change (see eqn. 1) and it should be related to
ofthemicroquasar
ACKNOWLEDGMENTS
We acknowledge the contributions of the scientific and tech-
nical staff of TIFR, ISAC and ISTRAC for the successful
fabrication, launch and operation of the IXAE. We thank
Dr. S. Seetha and Dr. K. Kasturirangan for their contribu-
tion to the IXAE. It is a pleasure to acknowledge constant
support of Shri K. Thyagarajan, Project Director IRS−P3
satellite, Shri R. N. Tyagi, Project Director IRS−P4 satel-
lite, Shri J. D. Rao and his team at ISTRAC, Dr P. S.
Goel, Director ISAC and the Director of the ISTRAC. SN
thanks M. Choudhury for his comments and discussions
on the manuscript. This research has made use of data
obtained through the High Energy Astrophysics Science
Archive Research Center Online Service, provided by the
NASA/Goddard Space Flight Center. Finally we thank an
anonymous referee for his comments and suggestions which
resulted in very significant improvement of this paper.
REFERENCES
Agrawal, P. C., Perspective in High Energy Astronomy and As-
trophysics; Proceedings of the International Colloquium, Aug.
c ? 0000 RAS, MNRAS 000, 000–000

Page 10

10
Naik, S. et al.
12-17, 1998 at TIFR, Mumbai, eds: P. C. Agrawal and P. R.
Viswanath, University Press, Hyderabad, India, p408
Belloni, T., Klein-Wolt, M., Mendez, M., van der Klis, M., and
van Paradijs, J. 2000, A&A, 355, 271
Castro-Tirado, A.J., Brandt, S., & Lund, N. 1992, IAU Circ.,
5590.
Chakrabarti, S. K. 1998, in Observational Evidence for Black
Holes in the Universe, ed. S. K. Chakrabarti (Dordrecht:
Kluwer), 19
Chakrabarti, S. K. & Manickam, S. G. 2000, ApJ, 531, L41
Chakrabarti, S. K., Manickam, S. G., Nandi, A., Rao, A. R. 2000,
in Ruffini R., ed., Proc. 9th Marcel Grossman Conf. World
Scientific, Singapore (in press)
Greiner, J., Morgan, E. H., & Remillard, R. A. 1998, New Astron.
Rev., 42, 597
Mirabel, I. F. & Rodriguez, L. F. 1994, Nature, 371, 46
Morgan, E. H., Remillard, R. A., & Greiner, J. 1997, ApJ, 482,
993
Muno, M. P., Morgan, E. H., & Remillard, R. A. 1999, ApJ, 527,
321
Naik, S., & Rao, A. R. 2000, A&A, 362, 691
Naik, S., Agrawal, P. C., Rao, A. R., Paul, B., Seetha, S., &
Kasturirangan, K. 2001, ApJ, 546, 1075
Paul, B., Agrawal, P. C., Rao, A. R., Vahia, M. N., Yadav, J. S.,
Seetha, S., & Kasturirangan, K. 1998, ApJ, 492, L63
Rao, A. R., Agrawal, P. C., Paul, B., Vahia, M. N., Yadav, J. S.,
Marar, T. M. K., Seetha, S., & Kasturirangan, K. 1998, A&A,
330, 181
Rao, A. R., Yadav, J. S., & Paul, B. 2000, ApJ, 544, 443
Taam, R. E., Chen, X., & Swank, J. H. 1997, ApJ, 485, L83
Vilhu, O., & Nevalainen, J. 1998, ApJ, 508, L85
Yadav, J. S., Rao, A. R., Agrawal, P. C., Paul, B., Seetha, S., &
Kasturirangan, K. 1999, ApJ, 517, 935
c ? 0000 RAS, MNRAS 000, 000–000