ap-6-11.dvi


Acta Polytechnica Vol. 51 No. 6/2011

Monitoring PSR B1509–58 with RXTE: Spectral analysis 1996–2010

E. Litzinger, K. Pottschmidt, J. Wilms, S. Suchy, R. E. Rothschild, I. Kreykenbohm

Abstract

We present an analysis of the X-ray spectra of the young, Crab-like pulsar PSR B1509–58 (pulse period P ∼ 151ms)
observed by RXTE over 14 years since the beginning of the mission in 1996. The uniform dataset is especially well suited
for studying the stability of the spectral parameters over time as well as for determining pulse phase resolved spectral
parameters with high significance. The phase averaged spectra as well as the resolved spectra can be well described by
an absorbed power law.

Keywords: pulsars: individual (PSR B1509–58) — stars: neutron — X-rays: stars.

1 Introduction
The pulsar PSR B1509–58 was discovered in Einstein
X-Ray Observatory data from 1979 and 1980 [5].
The pulsar is associated with the supernova remnant
G320.4–1.2 (MSH 15–52) in the constellation Circi-
nus.
PSR B1509–58 has been established as one of only

a few known Crab-like sources, i.e., a young pulsar
powering a synchrotron nebula [6]. PSR B1509–58’s
nebula is considerably larger, its surface brightness
is lower and the pulse period of P ∼ 151ms is
slower than that of the Crab. Due to a very high
spin-down rate of Ṗ ∼ 1.5 × 10−12 s s−1, however,
the characteristic age P/2Ṗ of PSR B1509–58 is
∼ 1.6 × 103 yr (e.g., [7]), i.e., comparable to that of
the Crab (∼ 1.3× 103 yr).
In the following we analyse the phase averaged

spectra (Sec. 2) and the phase resolved spectra
(Sec. 5) from calculated ephemerides (Sec. 3) for the

three major calibration epochs 3–5 for the RXTE.
Pulse profiles for the epochs are presented in Sec. 4.
A short summary of the results and the implications
for further spectral analysis with RXTE are given in
Sec. 6

2 Phase averaged spectra

From approximately monthly monitoring observa-
tions of PSR B1509–58 time averaged PCA (PCU2
top layer, [1, 2]) and HEXTE (cluster A and B, [4])
spectra were created by averaging individual mon-
itoring spectra over major instrument calibration
epochs, i.e. between MJD 50188 and 51259 (epoch 3),
MJD 51259 and 51677 (epoch 4), and from MJD
51677 onward (epoch 5). Figure 1 shows the epoch
averaged counts spectra for epoch 5. Some spectral
parameters and their uncertainties for the three fits
are given in Table 1. No systematic uncertainties
have been added to the spectra.

Fig. 1: (a) PCU2 top layer and HEXTE counts spectra obtained by accumulating all suitable monitoring spectra of
epoch 5 and best simultaneous fit model (absorbed power law with an iron line), (b) Best fit residuals

38



Acta Polytechnica Vol. 51 No. 6/2011

Table 1: Best fit parameters for the phase averaged PCU2 spectra of calibration epoch 3–5

Parameter Epoch 3 Epoch 4 Epoch 5

Γ 2.022± 0.001 2.021± 0.001 2.026± 0.001
AΓ

[
10−2 keV−1 cm−2 s−1

]
7.48± 0.84 7.48± 1.01 7.33± 0.12

NH
[
1022 cm−2] 0.37± 0.01 0.39± 0.02 0.58± 0.02

EFe [keV] 6.65+0.01−0.16 6.50+0.16−0.03 6.50± 0.03
cHEXTE 0.65± 0.05 0.76± 0.04 0.85± 0.02
χ2red/dof 1.09/48 0.58/42 2.16/94

F4−10 keV
[
10−11 erg cm−2s−1

]
10.25 10.34 9.85

F unabs4−10 keV
[
10−11 erg cm−2s−1

]
10.57 10.57 9.93

F10−20 keV
[
10−11 erg cm−2s−1

]
7.76 7.76 7.20

F20−200 keV
[
10−11 erg cm−2s−1

]
25.07 25.15 22.36

The spectra are modeled by an absorbed power
law. In addition we found clear indications of a
narrow iron Kα line which is included by a Gaus-
sian component added to the power law. Be-
cause of the long monitoring time of epoch 5
(ten years) systematic features are visible in the
spectra. For the PCA there are strong residu-
als around 5 keV depending on the three xenon L
edges. Around 9.3 keV a broad negative residual
is visible. We model it with an additional neg-
ative Gaussian line at this energy with a flux of
−3 × 10−5 cm−2 s−1. We speculate that this resid-
ual might be related to imperfect modeling of the
copper Kα emission line at 8.04 keV. Also events
from the americium calibration source of the PCA
at 33 keV are visible. They were also modeled with

a negative Gaussian line with a flux of −1.4 ×
10−4 cm−2 s−1.

3 Pulse period ephemeris

The pulse phase resolved analysis for the PCA
is based on high time-resolution GoodXenon event
mode data, filtered for PCU2 top layer events.
Ephemerides for PSR B1509–58 were calculated from
the pulse frequencies of each observation. The ref-
erence epoch was set to t0(MJD) = 52 921.0, the
averaged time of the monitoring (see Figure 2, Ta-
ble 2). With this result barycentered pulse phase-
and energy resolved source count rates (pha2 files)
were created using a modified version of the FTOOL
fasebin [3].

Fig. 2: Frequencies of each observation, calculated by epoch folding the barycentered GoodXenon lightcurve, vs. time
with the best fit of a polynomial of quartic grade. A linear decline is visible

Table 2: Values for the pulse frequency and its first three derivatives

ν
[
s−1

]
ν̇

[
10−11 s−2

]
ν̈

[
10−21 s−3

]
...
ν

[
10−29 s−4

]

6.610 32± 1e− 6 −6.696 5± 0.002 1.18± 0.24 1.77± 0.36

39



Acta Polytechnica Vol. 51 No. 6/2011

4 Pulse profiles

Pulse profiles in the energy range 3–43keV for the
three major epochs are shown in Figure 3. The peak
was centered to phase 1.0 by shifting the individ-
ual pulse profiles and adding them up. The decline
in rate between epochs is an instrumental effect and
is accounted for in the calibration of the PCU2 top
layer. The bigger errorbars in epoch 4 are due to
a shorter duration and therefore less observations
(30 in epoch 3, 13 in epoch 4, 213 in epoch 5). A
clear division in peak Φ = 0.88 − 1.25 and off-peak
Φ = 0.44− 0.75 is possible.

Fig. 3: Pulse profiles for the energy range 3–43 keV of
PSR B1509–58 for the three major epochs 3, 4 and 5

5 Phase resolved spectra

All pulsed (i.e., peak minus off-peak) and unpulsed
(regular background) spectra were summed to obtain
averaged spectra for epochs 3–5, respectively. For
the averaged phase resolved spectra, as for the phase
averaged spectra before, an absorbed power law was
fitted (energy range 3–20keV). For epoch 5 it showed
that the residuals for the pulsed emission improved
by including a cutoff (see Figures 4 and 5). The

pulsed spectra showed no indication for an iron line
at 6.4 keV. The line is part of the unpulsed spectra
and hence of the PWN of PSR B1509–58. This is also
the explanation for the different values of the NH. In
the pulsed emission we see the beam of the pulsar and
the galactic extinction. In the unpulsed emission we
see the surrounding PWN without the pulsar and the
galactic extinction. Therefore the value for the latter
is smaller than for the pulsed emission. The best fit
parameters for epoch 5 are shown in Table 3.

Fig. 4: PCU2 top layer counts spectra of the pulsed emis-
sion of epoch 5. Residuals are shown for an absorbed
power law and after including a cutoff

Fig. 5: The same as Figure 4 but for the unpulsed emis-
sion

Table 3: Best fit parameters for the pulsed and unpulsed emission for epoch 5

Parameter E5 pulsed E5 pulsed E5 unpulsed

Γ 1.37± 0.01 1.27± 0.01 2.27± 0.01
Ecut [keV] 116± 8
NH

[
1022cm−2] 2.40± 0.1 1.97± 0.11 1.34± 0.02

EFe [keV] 6.57+0.01−0.07

F4−10 keV
[
10−11 erg cm−2s−1

]
4.28± 0.02 4.29± 0.02 7.99± 0.01

F unabs4−10 keV
[
10−11 erg cm−2s−1

]
4.57± 0.02 4.52± 0.02 8.30± 0.01

F10−20 keV
[
10−11 erg cm−2s−1

]
5.56± 0.02 5.57± 0.02 4.90± 0.01

χ2red 1.09 0.91 1.26

40



Acta Polytechnica Vol. 51 No. 6/2011

6 Summary and conclusions

We could well describe the spectra with an absorbed
power law with a Gaussian line for the phase averaged
and the unpulsed spectra and without the Gaussian
for the pulsed spectra. For the pulsed spectrum of
epoch 5 a cutoff improves the fit, while for the other
epochs and the unpulsed emission it has no effect. No
significant changes between the values of the different
epochs for the averaged, pulsed and unpulsed emis-
sion were seen. Long observations show systematic
effects from the instruments and are therefore good
for describing the calibration effects. As forthcoming
work we intend to add HEXTE spectra to the epoch
averaged phase resolved analysis.

References

[1] Jahoda, K., Swank, J. H., Giles, A. B., et al.:
Proc. EUV, X-Ray, and Gamma-Ray Instrumen-
tation for Astronomy VII, 1996, Vol. 2 808, 59.

[2] Jahoda, K., Markwardt, C. B., Radeva, Y., et al.:
ApJS, 163, 401, 2006.

[3] Kreykenbohm, I., Coburn, W., Wilms, J., et al.:
A&A 395, 129, 2002.

[4] Rothschild, R. E., Blanco, P. R., Gruber, D. E.,
et al.: ApJ, 496, 538, 1998.

[5] Seward, F. D., Harnden, Jr. F. R.: ApJ, 256, L45,
1982.

[6] Seward, F. D., Harnden, Jr. F. R., Szymko-
wiak, A., et al.: ApJ, 281, 650, 1984.

[7] Zhang, L., Cheng, K. S.: A&A, 363, 575, 2000.

E. Litzinger
Dr. Remeis-Observatory/ECAP, FAU
Bamberg, Germany

K. Pottschmidt
CRESST/NASA-GSFC
Greenbelt, MD, USA
UMBC
Baltimore, MD, USA

J. Wilms
Dr. Remeis-Observatory/ECAP, FAU
Bamberg, Germany

S. Suchy
CASS/UCSD
La Jolla, CA, USA

R. E. Rothschild
CASS/UCSD
La Jolla, CA, USA

I. Kreykenbohm
Dr. Remeis-Observatory/ECAP, FAU
Bamberg, Germany

41