Free Access
Volume 579, July 2015
Article Number A22
Number of page(s) 2
Section Extragalactic astronomy
Published online 19 June 2015

© ESO, 2015

1. Introduction

Ultraluminous X-ray sources (ULX) are empirically defined as non-nuclear extragalactic sources with luminosities exceeding ~1039 erg s-1, i.e. the Eddington limit for isotropic accretion on stellar-mass black holes (Feng & Soria 2011). These sources are considered, therefore, the best candidates to host intermediate mass black holes (IMBHs, Colbert & Mushotzky 1999). This interpretation is also supported by observations of quasi-periodic oscilations in several ULXs (Strohmayer & Mushotzky 2009), which might imply blackhole masses exceeding several thousands of solar mass. However, interpretation is not straighforward (Middleton et al. 2011) and super-Eddington accretion onto stellar mass black holes like in the case of the well-known source SS 433 (King et al. 2001; Fabrika 2004) might still be an option. Emission from young rotation-powered pulsars (Medvedev & Poutanen 2013) has also been suggested as a possible mechanism powering ULXs.

Recently, the discovery of coherent X-ray pulsations from ULX M82 X-2 (or X-1) (Bachetti et al. 2014), with a period of ~ 1.37 s and the associated orbital motion, provided a new explanation for the very bright emission of at least some of these objects. Based on the X-ray timing Bachetti et al. (2014) have derived a mass function of f = 2.1 M. The accretion would then proceed from a low-mass companion via Roche-lobe overflow, likely the only mechanism to feed enough matter to explain the observed luminosity. This implies a strong spin-up torque imposed onto the neutron star in agreement with the observed short spin-up timescale (P/ ~ 300 yr) and spin period of the pulsar. In principle, other ULXs powered by an accreting neutron star could exist and are expected to have spin periods of the order of 10 s as well.

Currently there is a lack of detailed studies of the variability in ULXs at short timescales (i.e. 100 s) as illustrated by the surprising discovery of the pulsations in M82 X-2, which is a rather well-studied system. Motivated by this, we performed a targeted search for coherent pulsations with periods in the range 0.15–15 s in archival XMM-Newton data to understand whether other ULXs might host accreting pulsars. In the present note, we report the results of this systematic search.

2. Source sample and available data

The number of ULXs and candidates has been steadily increasing since the launch of the Einstein satellite (Feng & Soria 2011). Many of them are, however, not sufficiently bright for detailed timing analysis. As an initial step, we limited our analysis to a sample of fifteen bright ULXs observed with XMM-Newton. This sample was used by Heil et al. (2009) to characterize ULX variability in a broad frequency range and no significant periodicities were reported by these authors. However, Heil et al. (2009) did not search for coherent pulsations.

We searched for pulsation in archival XMM-Newton data of the EPIC PN camera as it is one of the few instruments with adequate timing resolution to investigate variability down to subsecond timescales. Our analysis is based on a larger dataset compared to the one considered by Heil et al. (2009) since many additional observations have become available since their publication. Most of the observations were performed in full-frame read-out mode with time resolution of ~ 0.07 s. The list of the sources in our sample, and the summary of available observations are presented in Table 1.

Table 1

Sources included in the analysis.

3. Data analysis and results

Low-level data reduction was carried out using the XMM SAS 13.5 package, current calibration files, and standard filtering criteria1. Periods of high background flaring activity were filtered using the espfilt task based on the observed count-rate of the non-exposed detector corners and visual inspection of the detector lightcurves above 10 keV for periods of high background. Source photons with energies in 0.3–10 keV range were then extracted using the source-centred circles with radius of 20′′ in all cases except M82 X-2, where to match the original NuStar observation more closely we used a radius of 40′′ centred in between the two nearby ULXs, X-1 and X-2, where the pulsations could potentially originate from (Bachetti et al. 2014). We note that even with a larger extraction region the observed power spectrum remains dominated around pulsation frequency by the white noise due to the limited photon statistics, so increasing the extraction region size is not expected to change the sensitivity to the coherent signal.

The arrival times of individual photons were then corrected to solar barycentre using the barycen task. To search for pulsations, we used the H-test (de Jager et al. 1989) applied to unbinned source events. Taking into account that in M82 X-2 the pulsed fraction showed large variations with time (suggesting the possible transient nature of the pulsations), we analysed individual observations separately. Our results are summarized in Table 1.

We have found no evidence for significant periodic signals in any of the sources. Therefore, we followed the approach suggested by Brazier (1994) for deriving the upper limits on amplitude of potentially present but undetected periodic signals fpulsed. For each source and observation we calculated the upper limit on the amplitude of a sinusoidal signal with period in the range 0.15–15 s at a 3σ confidence level (for details see Eq. (3) and accompanying text in Brazier 1994). We report in Table 1 the highest upper-limit among all the observations available for a given source.

We note that for M82 X-2 our limit is below the lowest pulsation amplitude value reported by Bachetti et al. (2014). To illustrate this point, in Fig. 1 we present the periodogram for the longest XMM-Newton observation along with two periodograms for two simulated signals with amplitudes of 2.2% and 5% (corresponding to the derived upper limit and lowest amlitude reported by Bachetti et al. 2014, respectively). We note that a pulsed signal is not detected in XMM-Newton lightcurve and is easily detectable in simulated lightcurves, so the pulsations would be easily detectable if the pulsed fraction during the XMM-Newton observation was comparable to that reported for NuStar.

thumbnail Fig. 1

Periodogram for the longest observation of the pulsating in ULX in M82 (thick red line, obsid 0206080101, exposure of 46 ks), and periodograms for simulated sinusoidal signal with pulsed fractions of 5% (thin black line, only visible at the peak) and 2.2% (dotted green line) with the same exposure, observation duration, and total number of photons as observed.

Open with DEXTER

4. Conclusions

Inspired by the recent discovery of the pulsations from the M82 X-2 with NuStar, we revisited the available archival XMM-Newton observations of several bright ULXs in order to systematically search for pulsations whose detection escaped previous investigations. We found no significant pulsed signal in the range of periods from 0.15 to 15 s in any of the considered sources including M82 X-2. We provide, therefore, upper limits for pulsed fraction of potentially non-detected pulsations. We note that in many cases these limits are rather weak owing to limited statistics and could be significantly improved with additional observations.

For M82 X-2 our upper limit turns out to be a factor of two lower than the lowest value fpulsed ~ 5% reported by Bachetti et al. (2014), and therefore, we exclude pulsations with amplitude similar to that observed by NuStar in the XMM-Newton data. The amplitude of the pulsations in this source, however, has been reported to vary with time, hence it cannot be excluded that at the time of the XMM-Newton observation it remained intrinsically low. This would imply that pulsations in ULXs powered by accreting neutron stars might be transient and highlights the importance of regular monitoring of ULXs, particularly at higher energies where the pulsed fraction is expected to be larger. Still, an independent confirmation of pulsations in M82 X-2 would be indispensable.


The authors thank the Deutsches Zentrums für Luft- und Raumfahrt (DLR) and Deutsche Forschungsgemeinschaft (DFG) for financial support (grants DLR 50 OR 0702, FKZ 50 OG 1301, SA2131/1-1).


  1. Bachetti, M., Harrison, F. A., Walton, D. J., et al. 2014, Nature, 514, 202 [NASA ADS] [CrossRef] [Google Scholar]
  2. Brazier, K. T. S. 1994, MNRAS, 268, 709 [NASA ADS] [Google Scholar]
  3. Colbert, E. J. M., & Mushotzky, R. F. 1999, ApJ, 519, 89 [NASA ADS] [CrossRef] [Google Scholar]
  4. de Jager, O. C., Raubenheimer, B. C., & Swanepoel, J. W. H. 1989, A&A, 221, 180 [NASA ADS] [Google Scholar]
  5. Fabrika, S. 2004, Astrophys. Space Phys. Rev., 12, 1 [NASA ADS] [Google Scholar]
  6. Feng, H., & Soria, R. 2011, New A Rev., 55, 166 [NASA ADS] [CrossRef] [Google Scholar]
  7. Heil, L. M., Vaughan, S., & Roberts, T. P. 2009, MNRAS, 397, 1061 [NASA ADS] [CrossRef] [Google Scholar]
  8. King, A. R., Davies, M. B., Ward, M. J., Fabbiano, G., & Elvis, M. 2001, ApJ, 552, L109 [NASA ADS] [CrossRef] [Google Scholar]
  9. Medvedev, A. S., & Poutanen, J. 2013, MNRAS, 431, 2690 [NASA ADS] [CrossRef] [Google Scholar]
  10. Middleton, M. J., Roberts, T. P., Done, C., & Jackson, F. E. 2011,MNRAS, 411, 644 [Google Scholar]
  11. Strohmayer, T. E., & Mushotzky, R. F. 2009, ApJ, 703, 1386 [NASA ADS] [CrossRef] [Google Scholar]

All Tables

Table 1

Sources included in the analysis.

All Figures

thumbnail Fig. 1

Periodogram for the longest observation of the pulsating in ULX in M82 (thick red line, obsid 0206080101, exposure of 46 ks), and periodograms for simulated sinusoidal signal with pulsed fractions of 5% (thin black line, only visible at the peak) and 2.2% (dotted green line) with the same exposure, observation duration, and total number of photons as observed.

Open with DEXTER
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.