DISCOVERY OF AN ENERGETIC 38.5 ms PULSAR POWERING THE GAMMA-RAY SOURCE IGR J18490−0000/HESS J1849−000

The detection of 10¹² eV emission associated with supernova products in the Galaxy has opened up a new window on the evolution of these energetic stellar remnants. More than 2/3 of the now >60 Galactic TeV sources are supernova remnants or pulsar wind nebulae (PWNe), the latter being the largest class.³ Many of these TeV PWNe are spatially offset from middle-aged (τ_c ∼ 10⁴–10⁵ year old) pulsars and are often more extended and more luminous than their X-ray PWNe counterparts. This TeV emission is likely to be inverse Compton scattered radiation from relic electrons produced by the pulsar in an earlier stage of energetic spin-down (de Jager & Djannati-Ataï 2008; Zhang et al. 2008; Mattana et al. 2009b). In contrast, younger (τ_c ∼ 10³ year old) pulsars are associated with compact TeV sources that are co-located with their X-ray PWNe. In these younger systems the high magnetic fields make for efficient synchrotron X-ray sources and inefficient inverse Compton TeV emission.

HESS J1849−000 is one of the fainter sources detected in the HESS Galactic Plane survey (Aharonian et al. 2005, 2006) and subsequent dedicated observations, with a significance of 6.4σ, flux >350 GeV of ≈15 mCrab, and only weak indication of extent (Terrier et al. 2008). HESS J1849−000 is coincident in position with the INTEGRAL/IBIS source IGR J18490−0000 that was discovered during a survey of the tangent regions of the Sagittarius and Scutum spiral arms (Molkov et al. 2004; Bird et al. 2006). Follow-up X-ray observation of IGR J18490−0000 by Rodriguez et al. (2008) with the Swift X-ray telescope (XRT) located a highly absorbed X-ray point source within the INTEGRAL/IBIS error circle. These authors suggested a Galactic X-ray binary origin for IGR J18490−0000 based on a K-band Two Micron All Sky Survey star located within the Swift XRT error circle. However, that association was excluded by the precise X-ray position obtained in a brief Chandra High Resolution Camera (HRC) observation (Ratti et al. 2010). Terrier et al. (2008) used imaging and spectroscopic evidence from an XMM-Newton observation to argue that IGR J18490−0000 is instead a pulsar/PWN. As such, it would join the dozen hard X-ray members of this class detected by INTEGRAL/IBIS (Mattana et al. 2009a; Renaud et al. 2010).

In Section 2, we present in detail the XMM-Newton imaging and spectral data that support a pulsar/PWN interpretation for IGR J18490−0000. In Section 3, we report the discovery using Rossi X-ray Timing Explorer (RXTE) of PSR J1849−0001 and its spin-down, which verifies this conjecture. In Section 4, we discuss the properties of HESS J1849−000 in the context of the spin-down parameters of PSR J1849−0001.

An 11 ks XMM-Newton observation of IGR J18490−0000 (ObsID 0306170201) was acquired on 2006 April 3 using the European Photon Imaging Camera (EPIC; Turner et al. 2003). EPIC consists of three sensors operating in parallel, one pn and two MOS CCD cameras. These instruments are sensitive to X-rays in the 0.2–12 keV range with energy resolution $\Delta E/E \approx 0.1/\sqrt{E({\rm keV})}$. All three cameras were operated in full-frame mode, using the thin and medium filters for the EPIC pn and MOS, respectively. The target was placed at the default EPIC pn focal plane location for a point source. The time resolutions of 73.4 ms and 2.7 s for the EPIC pn and MOS, respectively, are insufficient to search for a typical pulsar signal.

The data were processed using the SAS version xmmsas_20061026_1802-6.6.0 pipeline and were analyzed using both the SAS and FTOOLS software packages. The observation was free of significant particle background contamination and provided a near continuous 11 ks of good observing time for the EPIC MOS and 9.9 ks for EPIC pn.

Figure 1 (left) displays the 2–10 keV XMM-Newton EPIC MOS X-ray image centered on IGR J18490−0000. The image has been exposure corrected and smoothed using a σ = 38 Gaussian kernel and scaled to highlight the faint diffuse emission near the central bright source at coordinates (J2000.0) R.A. = 18^h49^m0162, decl. = −00°01'177. This source, at the position of the previously detected Swift XRT source, lies at the center of the 14 radius error circle of IGR J18490−0000. There are no other prominent X-ray sources in the full 30' diameter field of this instrument. Together with the positional coincidence, its flux and spectrum (see below) leave no doubt that XMMU J184901.6−000117 is the counterpart of IGR J18490−0000. Faint, diffuse emission surrounds the point source, and a prominent feature extends up to 2' to the southwest; these properties are suggestive of a PWN. A 1.2 ks Chandra HRC image obtained on 2008 February 16 (Ratti et al. 2010) shows that most of the XMM-Newton source flux is contained in a point-like component.

Zoom In Zoom Out Reset image size

2.1. XMM-Newton Point-source Spectrum

2.2. XMM-Newton Extended Emission

To quantify the spatial structure of the XMM-Newton source, we made a radial profile of the counts distribution centered on the point source and compared it with the point-spread function (PSF). The latter was constructed using instrumental profiles at several energies weighted to match the measured spectral index. Figure 2 shows the measured EPIC pn profile compared to the average PSF for this spectrum and location in the field. Background has been taken from a symmetric position with respect to the optical axis. There is a clear excess of emission from 20'' to 150'' radius. We fitted the radial counts distribution to the sum of a King profile to represent the point source, and a Gaussian to represent the extended emission. A Gaussian component 150'' ± 15'' in diameter improves the quality of the fit by ≈9σ, proving that the source is surrounded by a nebula or a halo.

Zoom In Zoom Out Reset image size

We extracted the spectrum of the extended nebula in an annulus surrounding XMMU J184901.6−000117 from 30'' to 150''. To account for contamination from the wings of the bright point source, we fitted the point source and extended emission spectra simultaneously. We used separate power laws but the same N_H for both spectra. We added a constant fraction of the point-source spectrum to the diffuse emission. This fraction was estimated from the PSF as 12% and 14% for EPIC pn and MOS, respectively, at a radius of 30''. To account for the uncertainty in the estimate of these fractions, we varied them from 7% to 19% of the point-source flux, adding the corresponding systematic uncertainty to errors on the fitted parameters. The resulting diffuse emission spectrum has Γ_PWN = 2.1 ± 0.3 and a flux of F_PWN(2–10 keV) = (9 ± 2) × 10⁻¹³ erg cm⁻² s⁻¹. The reduced chi-squared is χ²_ν = 0.90 for 311 dof.

The extended nebula with a steeper spectrum suggests that IGR J18490−0000 is a pulsar/PWN system. Alternatively, given the large N_H, some of the diffuse emission could be a dust-scattered halo of the bright point source (Predehl & Schmitt 1995); however that could not account for its asymmetric extension. We note that the nebula is rather faint, ≈25% of the point-source flux, but it is still in the range observed for PWNe (Kargaltsev & Pavlov 2008).

To search for the expected pulsar signal from IGR J18490−0000, we obtained 112.3 ks of RXTE observations, pointed at the Chandra source position, spanning 2010 November 25–December 15. The schedule was designed to obtain a phase-connected timing solution that would measure both P and $\dot{P}$. The data used here were collected with the Proportional Counter Array (PCA; Jahoda et al. 1996) in the GoodXenon mode with an average of 1.75 out of the five proportional counter units (PCUs) active. In this mode, photons are time-tagged to 0.9 μs precision and have an absolute time accuracy of <100 μs (Rots et al. 1998). The effective area of five combined detectors is ≈6500 cm² at 10 keV with a roughly circular field of view of ∼1° FWHM. Spectral information is available in the 2–60 keV energy band with a resolution of ∼16% at 6 keV.

Standard time filtering was applied to the PCA production data, rejecting intervals of South Atlantic Anomaly passage, Earth occultation, and other periods of high particle activity. The photon arrival times were transformed to the solar system barycenter in Barycentric Dynamical Time (TDB) using the JPL DE200 ephemeris and the coordinates given in Table 1, obtained from Chandra HRC observation ObsID 7398. These coordinates are derived from a centroid calculation of 17 photons within a 2'' radius aperture and have nominal uncertainty of 06. We note that these coordinates differ by 035 from those reported in Ratti et al. (2010).

Table 1. Timing Parameters of PSR J1849−0001

Parameter	Value
R.A. (J2000.0)a	18h49m0161
Decl. (J2000.0)a	−00°01'176
Epoch (MJD TDB)	55535.285052933
Periodb, P	38.518943432(12) ms
Period derivativeb, $\dot{P}$	1.4224(58) × 10−14 s s−1
Range of dates (MJD)	55525.6–55545.2
Spin-down luminosity, $\dot{E}$	9.8 × 1036 erg s−1
Characteristic age, τc	42.9 kyr
Surface dipole magnetic field, Bs	7.5 × 1011 G

Notes. ^aChandra position with a nominal uncertainty of 06 (see the text). ^bTEMPO 1σ uncertainties given in parentheses.

Download table as:  ASCIITypeset image

3.1. RXTE Timing Analysis

We restricted the analysis to the 2–20 keV energy range (PCA channels 2–50) from the top Xenon layer of each PCU to optimize the signal-to-noise. A fast Fourier transform of the initial data set (ObsID 95309-01-01-02; 7 ks) revealed a highly significant signal of period P = 38.52 ms. Subsequently, for each satellite orbit we extracted a pulse profile corresponding to the peak power as determined by the Rayleigh test (Strutt 1880), also known as Z²₁ (Buccheri et al. 1983). The resulting profiles were cross-correlated, shifted, and summed to generate a master pulse profile template. Each individual profile was then cross-correlated with this template to determine its time of arrival (TOA) and uncertainty. These TOAs were then iteratively fitted to a quadratic ephemeris.

The resulting unique ephemeris is presented in Table 1, and the phase residuals are shown in Figure 3. The residuals are all less than 0.05 cycles and do not require a higher derivative. The derived properties of PSR J1849−0001 are: spin-down luminosity $\dot{E} = 4\pi ^2I\dot{P}/P^3 =$ 9.8 × 10³⁶ erg s⁻¹, characteristic age $\tau _c \equiv P/2\dot{P} =$ 42.9 kyr, and surface dipole magnetic field strength $B_s = 3.2 \times 10^{19}(P\dot{P})^{1/2} =$ 7.5 × 10¹¹ G.

Zoom In Zoom Out Reset image size

Figure 3 displays the summed pulse profile using all the 2–20 keV data folded on the final ephemeris. It has a symmetric, single-peaked structure. The raw pulsed fraction is 3.8%, uncorrected for PCA instrument or astrophysical background, difficult to determine with sufficient accuracy for this purposes. We see no energy dependence of the pulse profile when subdividing the 2–20 keV band.

3.2. RXTE Spectral Analysis

The spectrum of the pulsed flux from PSR J1849−0001 can be isolated by phase-resolved spectroscopy. We used the fasebin software to construct phase-dependent spectra based on the ephemeris of Table 1. For each ObsID and PCU we constructed spectra from counts detected in the top Xenon layer only and combined them to produce a single spectrum per PCU for the entire set of observations. Similarly, standard PCA responses for each PCU were generated at each ObsID and averaged. In fitting the pulsed flux, the unpulsed emission provides a near perfect background estimate and was taken from the 0.4 phase bins range corresponding to the region of flat minimum in the pulse profile around phase zero.

The merged RXTE spectrum was fitted in the 2–20 keV range using a simple absorbed power-law model with the interstellar absorption held fixed at N_H = 4.3 × 10²² cm⁻² determined from the XMM-Newton fit (see Figure 4). The resulting best-fit photon index is Γ = 1.3^+0.2_−0.3. The 2–10 keV pulsed flux is 8.9 × 10⁻¹³ erg cm⁻² s⁻¹, which represents ∼25% of the point-source flux of PSR J1849−0001 measured by XMM-Newton. The corresponding pulsed luminosity (assumed isotropic) is 6.7 × 10³³ d²₇ erg s⁻¹. The reduced chi-squared is χ²_ν = 0.70 for 29 dof.

Zoom In Zoom Out Reset image size

As expected, the RXTE measured spectral slope is intermediate between the XMM-Newton and INTEGRAL/IBIS values, indicating a steepening around 10–20 keV.

The discovery of PSR J1849−0001 using RXTE verifies the conjecture of Terrier et al. (2008) that XMMU J184901.6−000117, IGR J18490−0000, and HESS J1849−000 are all manifestations of a young pulsar/PWN system. Even though the RXTE field of view is wider than those of XMM-Newton and Chandra, there is little doubt that PSR J1849−0001 is the compact source in IGR J18490−0000, given the morphological and spectral evidence, and the compatibility of the RXTE measured pulsed flux with that of XMMU J184901.6−000117. Terrier et al. (2008) estimated that the spin-down luminosity of PSR J1849−0001 would be ≈9 × 10³⁶ erg s⁻¹ based on the empirical correlation between $\dot{E}$ and Γ_PSR of Gotthelf (2003); this turns out to have been an accurate prediction.

Assuming a distance of 7 kpc as proposed in Section 2.1, the X-ray luminosities of the pulsar and PWN, $L_{\rm PSR}(2\hbox{--}10\,{\rm keV}) = 2.9 \times 10^{-3}\,\dot{E}\,d_{7}^2$ and $L_{\rm PWN}(2\hbox{--}10\,{\rm keV}) = 6.6 \times 10^{-4}\,\dot{E}\,d_{7}^2$, respectively, are consistent with the range of pulsar/PWN systems (Kargaltsev & Pavlov 2008). The 20–100 keV flux of IGR J18490−0000 measured by INTEGRAL/IBIS (2 × 10⁻¹¹ erg cm⁻² s⁻¹; Terrier et al. 2008) corresponds to $L(20\hbox{--}100\,{\rm keV}) = 0.012\,\dot{E}\,d_{7}^2$. Given the 2–20 keV flux and spectral information from XMM-Newton and RXTE, the INTEGRAL/IBIS source is likely to be dominated by the pulsar rather than the PWN.

The TeV flux from HESS J1849−000 (2.2 × 10⁻¹² erg cm⁻² s⁻¹; Terrier et al. 2008) corresponds to $L(0.35\hbox{--}10\,{\rm TeV}) = 1.3 \times 10^{-3}\,\dot{E}\,d_{7}^2$. Such a small efficiency of converting spin-down power to high-energy radiation is typical of high $\dot{E}$ pulsars. The ratio F(0.35–10 TeV)/F_PWN(2–10 keV) ≈ 2 falls squarely on the inverse correlation between this quantity and $\dot{E}$ that was fitted by Mattana et al. (2009b), and modeled by them in terms of an evolving PWN emitting synchrotron X-rays and inverse Compton γ-rays. PSR J1849−0001 is in transition from a synchrotron-dominated X-ray PWN to an inverse Compton scattered TeV nebula, an expected phase through which a middle-aged pulsar will pass.

We thank the RXTE project for making the time available for this program, and the mission planners for carefully scheduling the observations. This investigation is also based on observations obtained with XMM-Newton, an ESA science mission with instruments, and contributions directly funded by ESA Member States, and NASA.