The New Magnetar SGR J1830−0645 in Outburst

The term "magnetar" was coined almost three decades ago to identify isolated neutron stars (NSs) ultimately powered by the dissipation of their own magnetic energy, which usually implies that they are endowed with huge magnetic fields, up to ∼10¹⁵ G (Duncan & Thompson 1992). A large fraction of the ∼30 magnetars known to date (Olausen & Kaspi 2014) have been discovered just over the past two decades, through their distinctive high-energy phenomenology: bursts of X-ray/gamma-ray emission and/or enhancements of their persistent X-ray luminosity, dubbed "outbursts" (see Kaspi & Beloborodov 2017; Esposito et al. 2021). The bursts are comparatively brief episodes lasting from milliseconds to hundreds of seconds, and reaching X-ray peak luminosities within the interval 10³⁹–10⁴⁷ erg s⁻¹ (e.g., Collazzi et al. 2015). The outbursts are instead long-lasting events where the X-ray luminosity first rises to values in the range of 10³⁴–10³⁶ erg s⁻¹, and then decreases on timescales that can be as long as years (Coti Zelati et al. 2018). Remarkably, in a couple of magnetars the post-outburst luminosity was found to differ from the pre-outburst long-term persistent level (Younes et al. 2017; Coti Zelati et al. 2020).

On 2020 October 10 at 14:49:24 UT, the Burst Alert Telescope (BAT) on board the Neil Gehrels Swift Observatory triggered on a short, hard X-ray burst (Page et al. 2020). Prompt observations with the Swift X-ray Telescope (XRT) localized a new uncatalogued X-ray source, SGR J1830−0645.¹⁴ An X-ray periodic signal at ∼10.4 s was detected in the XRT data (Gogus et al. 2020a) and confirmed later by observations with the Neutron Star Interior Composition Explorer (NICER) (Younes et al. 2020). The burst properties, the periodicity detected in the prompt follow-up observations, and the proximity of the source to the Galactic plane (Galactic latitude b ∼ 15) point to a newly discovered magnetar in outburst.

This Letter reports on (i) the properties of the X-ray bursts detected from SGR J1830−0645 by the Swift/BAT; (ii) quasi-simultaneous observations with XMM-Newton and the Nuclear Spectroscopic Telescope Array (NuSTAR) performed within ∼2 days after the first BAT burst; (iii) a Swift/XRT monitoring campaign covering the first month since the outburst onset; (iv) a search for short bursts in the X-ray time series (Section 2); and (v) radio observations with the Sardinia Radio Telescope and Parkes (Section 3). Discussion and conclusions follow (Section 4).

2.1. Observations and Data Analysis

2.2. Results

2.2.1. BAT Bursts Properties

Figure 1 shows the time evolution of the burst detected on 2020 October 10 in the 15–50 keV band (no emission is detected at higher energies). The event was single-peaked and had a total duration and a T₉₀ duration,¹⁶ as computed from a Bayesian blocks analysis with the battblocks tool, of 6 ± 1 ms and 5 ± 1 ms, respectively. The spectrum extracted from the whole interval, similarly to standard magnetar bursts seen at hard X-rays, can be described in the 15–50 keV band by simple models, such as a blackbody with temperature ${kT}={8.0}_{-1.2}^{+1.6}$ keV (with reduced χ² of ${\chi }_{r}^{2}=1.05$ for 14 degrees of freedom (d.o.f.)) or a power law with photon index Γ = 1.9 ± 0.5 (${\chi }_{r}^{2}=1.12$ for 14 d.o.f.). For the blackbody model, the average flux was ${7.6}_{-3.3}^{+0.4}\times {10}^{-7}$ erg cm⁻² s⁻¹, corresponding to a fluence of ∼4.6 × 10⁻⁹ erg cm⁻² (15–50 keV). Two other bursts resulted in BAT triggers during our campaign (see Palmer 2020). The second burst detected on 2020 November 5 was longer (25 ± 5 ms, T₉₀ = 21 ± 5 ms) and harder, being visible in the light curve up to ≈150 keV. Also in this case, the adoption of a blackbody model resulted in the best fit, with ${\chi }_{r}^{2}=0.83$ for 56 d.o.f., and kT = 9.8 ± 0.7 keV. The average flux was (1.3 ± 0.1) × 10⁻⁶ erg cm⁻² s⁻¹, while the fluence was ∼3.2 × 10⁻⁸ erg cm⁻² (15–150 keV). The burst that triggered BAT on 2020 November 11 was intermediate in hardness between the other two events, and lasted 26 ± 6 ms (T₉₀ = 21 ± 7 ms). For the blackbody model (kT = 9.1 ± 0.6 keV; ${\chi }_{r}^{2}=0.67$ for 56 d.o.f.), the average flux was (8.1 ± 0.7) × 10⁻⁷ erg cm⁻² s⁻¹ and the fluence ∼2.2 × 10⁻⁸ erg cm⁻² (15–150 keV).

Zoom In Zoom Out Reset image size

2.2.2. X-Ray Monitoring and Archival Observations

First, we fit the broadband spectrum extracted from the quasi-simultaneous EPIC-pn and NuSTAR data sets using models comprising different combinations of blackbody and power-law components. We included a constant term in the fits to account for inter-calibration uncertainties between the three instruments, deriving a mismatch of <5% for all the tested models. The best description of the data was provided by an absorbed double-blackbody model plus a power-law component accounting for emission at energies above ∼12 keV (Figure 1), giving ${\chi }_{r}^{2}$ = 1.10 for 376 d.o.f. and a null hypothesis probability nhp ≃ 0.10. Best-fitting parameters were absorption column density N_H = (1.07 ± 0.02) × 10²² cm⁻², blackbody temperatures and radii kT_BB,W = 0.45 ± 0.01 keV, R_BB,W = 5.6 ± 0.3 km for the warm component and kT_BB,H = 1.11 ± 0.01 keV, R_BB,H = 1.53 ± 0.03 km for the hot component, respectively (assuming a distance of 10 kpc; see Section 4), and Γ = 0.9 ± 0.3. The observed and unabsorbed fluxes were (4.33 ± 0.03) × 10⁻¹¹ and (5.54 ± 0.03) × 10⁻¹¹ erg cm⁻² s⁻¹ in the 0.3–25 keV energy range, with a fractional contribution of the power-law component of ≃6% in the same band.

Then, we fit an absorbed double-blackbody model to all Swift/XRT data, fixing the column density at N_H = 1.07 × 10²² cm⁻², and allowing all other parameters to vary across the data sets. The observed and unabsorbed fluxes derived from the above model are reported in Table 1, while the evolution of the unabsorbed flux is shown in the inset of the left panel in Figure 1. The unabsorbed flux decreased by a factor of ∼2 along our campaign, from ∼5 × 10⁻¹¹ erg cm⁻² s⁻¹ at peak to ∼2.5 × 10⁻¹¹ erg cm⁻² s⁻¹ about one month later (0.3–10 keV). Its time evolution can be adequately described so far by an exponential function with e-folding time τ = 55 ± 2 days (${\chi }_{r}^{2}$ = 1.42 for 13 d.o.f.; Figure 1).

SGR J1830−0645 was in the field of view of ROSAT/PSPC in a pointing performed on 1991 April 3 (see Table 1). The source is not detected, and we set an upper limit on the net count rate of 0.008 counts s⁻¹ (3σ c.l.; 0.1–2.4 keV). Assuming emission from the entire surface (R_NS = 10–15 km) and a source distance of 10 kpc, we estimate that the blackbody temperature should be ≲0.15 keV to be consistent with the above limit. The corresponding limits on the observed and unabsorbed flux are F_X,obs < 8 × 10⁻¹⁴ erg cm⁻² s⁻¹ and F_X,unabs < 1.5 × 10⁻¹² erg cm⁻² s⁻¹ (0.3–10 keV).

2.2.3. Timing Analysis and Phase-resolved Spectroscopy

The EPIC-pn, Swift/XRT, and NuSTAR source event files were used to study the magnetar timing properties. We built up a phase-coherent timing solution starting from the period P = 10.41572(1) s inferred from the EPIC-pn data sets (those with the largest statistics), and employing a phase-fitting technique. Within a baseline of about 34 days (2020 October 10–November 13), we clearly detected a first period derivative component in the signal phase history, and derived the following timing solution: P = 10.415724(1) s, $\dot{P}=7(1)\times {10}^{-12}$ s s⁻¹ at the reference epoch T₀ = 59133.0 MJD. The $\dot{P}$ value is in agreement with that derived by Ray et al. (2020) using NICER data sets. Our timing solution implies an rms variability of ∼145 ms, corresponding to a timing noise level <2%, similar to the range of values observed in other isolated NSs. We set a 3σ upper limit on the second period derivative of $| \ddot{P}| \lt 2\times {10}^{-17}$ s s⁻².

Figure 2 shows the background-subtracted light curves extracted from the EPIC-pn and NuSTAR data sets over different energy intervals, folded using the above ephemeris. Pulsed emission was detected up to an energy of ∼15 keV. The pulse profile displays an apparent complex morphology below 10 keV, with a pronounced dip close to the main peak, a weaker peak in the rising part of the profile, and small-amplitude structures at minimum (seemingly less prominent above ∼4 keV). The profile appears to evolve to a relatively simpler shape at higher energies in the 12–15 keV energy range. The pulse peak in this range lags the main peak observed at lower energies by 0.11 ± 0.06 spin phase cycles. The background-subtracted peak-to-peak semiamplitude increases from (63 ± 2)% below 3 keV to (71 ± 3)% in the 4–10 keV band, and drops to ∼20% in the 10–12 and 12–15 keV ranges. We set a 3σ upper limit of 20% in the 15–25 keV range (Figure 2). The marked changes in the pulse profile amplitude at energies where the power-law spectral component dominates the source emission (Figure 1) indicate that the power-law tail is also pulsed, though to a smaller extent than the low-energy blackbody components.

Zoom In Zoom Out Reset image size

We performed a phase-resolved spectral analysis over the 0.3–10 keV energy range using the EPIC-pn data. We extracted spectra from 50 phase intervals of width 0.02 rotational cycles, and fitted them using an absorbed double-blackbody model. In the fits, the column density was held fixed at the phase-averaged value (N_H = 1.07 × 10²² cm⁻²; Section 2.2.2), while all other parameters were allowed to vary. Figure 2 shows the evolution of the spectral parameters and unabsorbed fluxes of both thermal components, as well as their flux ratio, along the rotational phase. The complex spin modulation pattern for the soft X-ray emission can be ascribed to changes in the blackbody radius of both components. The smaller, hotter region traces more closely the fine structures seen in the pulse profile (e.g., the dip close to the peak; Figure 2). On the other hand, the scarce counting statistics available in the energy range 12–15 keV (less than 500 net counts summing up the NuSTAR FPMs) precludes an assessment of possible variability of the power-law slope along the rotational phase.

2.2.4. Search for Short X-Ray Bursts

Short X-ray bursts were searched for by applying the procedure described by Gavriil et al. (2004) and Scholz & Kaspi (2011; see also Borghese et al. 2020). We extracted light curves with different time resolutions (2⁻⁴, 2⁻⁵, 2⁻⁶ s) to improve sensitivity to bursts of different durations, except for Swift/XRT PC-mode light curves that were binned at the timing resolution (2.5073 s). For each observation, we calculated the Poisson probability of an event to be a random fluctuation with respect to the average number of counts per bin. Any bin with a probability smaller than ${10}^{-4}{({{NN}}_{\mathrm{trials}})}^{-1}$, where N is the total number of time bins and N_trials is the number of timing resolutions used in the search, was flagged as part of a burst. We detected 12 bursts in the Swift/XRT light curves (Figure 3 and Table 2). No bursts were found in the XMM-Newton and NuSTAR data sets.

Zoom In Zoom Out Reset image size

Table 2.  Log of X-Ray Bursts Detected in the Swift/XRT Data Sets

Obs.IDa	Burst Epoch	Durationb	Fluencec
	YYYY Mmm DD hh:mm:ss (TDB)	(ms)	(net counts)
00999571002 #1	2020 Oct 15 20:17:23	62.5	6
#2	20:21:01	62.5	6
#3	20:23:57	62.5	7
#4	20:27:01	62.5	11
#5	20:27:31	62.5	6
00999571004 #1	2020 Oct 19 13:56:24	62.5	15
#2	21:36:31	62.5	6
00013840001 #1	2020 Oct 29 17:26:40	62.5	7
#2	17:33:03	62.5	6
00013840004 #1	2020 Nov 6 11:39:58	62.5	6
00013840005 #1	2020 Nov 10 09:46:10	62.5	9
00013840006 #1	2020 Nov 13 01:43:12	62.5	7

Notes.

^aThe notation #N corresponds to the burst number in a given observation. ^bThe duration shall be considered as an approximate value. It was estimated as the coarser time resolution at which the burst is detected, or as the sum of the time bins showing enhanced emission for the case of the burst 00999571004 #1. ^cThe fluence refers to the 0.3–10 keV energy range.

Download table as:  ASCIITypeset image

Table 1 reports a log of the radio observations, performed using the Sardinia Radio Telescope (SRT; Bolli et al. 2015; Prandoni et al. 2017) and Parkes.

3.1. Sardinia Radio Telescope Observations

3.2. Parkes Observations

In 2020 October, SGR J1830−0645 entered its first detected outburst phase, revealing to be a new Galactic magnetar. The source emitted numerous X-ray bursts since the beginning of the outburst and during our monitoring campaign (Sections 2.2.1 and 2.2.4; see also Palmer 2020; Ray et al. 2020). This is a distinctive phenomenology usually observed in magnetars during an active phase. Our campaign allowed us to measure the spin period (P = 10.415724(1) s) and its first derivative ($\dot{P}=7(1)\times {10}^{-12}$ s s⁻¹) at the outburst peak (Section 2.2.3), hence to estimate a surface dipolar magnetic field ${B}_{\mathrm{dip}}\sim 6.4\times {10}^{19}{(P\dot{P})}^{1/2}\,\approx 5.5\times {10}^{14}$ G at pole (using the vacuum dipole formula), a spin-down luminosity ${\dot{E}}_{\mathrm{rot}}=4{\pi }^{2}I\dot{P}{P}^{-3}\approx 2.4\times {10}^{32}$ erg s⁻¹ (where I ≈ 10⁴⁵ g cm² is the moment of inertia of the NS) and a characteristic age ${\tau }_{c}=P/2\dot{P}\approx 24$ kyr.

The extension of the Galaxy in the direction of SGR J1830−0645 estimated from the maps of Hou & Han (2014) and the similarity between the column density derived from our spectral fits (N_H ∼ 1.07 × 10²² cm⁻²) and that expected in the direction of the source within the Galaxy (N_H,Gal ∼ 1.1 × 10²² cm⁻²; Willingale et al. 2013) suggest a distance D ≳ 5 kpc. In the following, we rescale all quantities to a distance of D = 10 kpc. Assuming isotropic emission, the X-ray luminosity at the outburst peak is then ${L}_{{\rm{X}},{\rm{p}}}\sim 6\times {10}^{35}{d}_{10}^{2}$ erg s⁻¹, while the limit we derived for the quiescent X-ray luminosity is ${L}_{{\rm{X}},{\rm{q}}}\lt 2\times {10}^{34}{d}_{10}^{2}$ erg s⁻¹ (0.3–10 keV; d₁₀ = D/10 kpc; see Section 2.2.2).

We studied the evolutionary history of SGR J1830−0645 using a two-dimensional magneto-thermal evolutionary model (Viganò et al. 2012, 2013, 2020). We used crustal-confined models consisting of an initial poloidal dipolar field (B_dip,in) plus a toroidal field (B_tor,in), and assumed for simplicity an equal amount of magnetic energy in the two components. We find that the evolution of an initial configuration with B_dip,in ∼ 10¹⁵ G and B_tor,in ∼ 10¹⁶ G provides a close match to the current P and $\dot{P}$ after ∼23 kyr. We thus obtain an age τ_th ∼ 23 kyr for SGR J1830−0645, similar to its characteristic age, and predict a quiescent bolometric thermal luminosity at τ_th that would be just below our current upper limit on L_X,q. The good agreement between τ_th and τ_c is due to the fact that the dissipation of the magnetic field is not yet substantial at this evolutionary stage. At later stages, τ_c will overestimate the real age because the electromagnetic torque decreases in time due to magnetic field dissipation, while τ_th usually remains consistent with the real age.

From its properties and simulated history, we find that SGR J1830−0645 is a middle age magnetar that had relatively strong magnetic energy at birth. The gray shaded region in Figure 4 shows the evolution of the spin period and luminosity of SGR J1830−0645 (taking into account uncertainties due to the assumption of light or heavy elements in the envelope), compared with current values for other magnetars. This region encompasses also two other magnetars with rather different magnetic field strengths,CXOU J171405.7−381031 and XTE J1810−197 (see the color scale in the figure). The properties of SGR J1830−0645 at birth were probably similar to those of CXOU J171405.7−381031, but different from those of XTE J1810−197. Indeed, CXOU J171405.7−381031 has a stronger magnetic field and a younger age than SGR J1830−0645, and it might be now at an earlier stage of a similar evolutionary scenario. On the other hand, XTE J1810−197 is characterized by a luminosity and spin period that might be potentially compatible with those expected if this source were at an earlier stage of the same evolutionary path of SGR J1830−0645, but its magnetic field is already smaller than that of SGR J1830−0645.

Zoom In Zoom Out Reset image size

The X-ray emission properties observed from SGR J1830−0645 soon after the outburst onset are in line with those of other magnetars (Coti Zelati et al. 2018 and references therein), and fit well within the resonant Compton scattering scenario (RCS; Thompson et al. 2002; Rea et al. 2008; Wadiasingh et al. 2018; see also Turolla et al. 2015 and references therein). The dominance of the thermal over the nonthermal (power-law) component (Section 2.2.2) suggests that the magnetospheric twist is restricted to a bundle of current-carrying field lines. Ohmic dissipation of the returning currents on the star surface leads to the appearance of localized hot spots (Beloborodov 2009). Heated regions may form on the surface also as a consequence of heat released deeper in the crust by local dissipation of magnetic energy (Pons & Rea 2012). In this respect, our analysis suggests the existence of two heated regions of different temperatures and sizes on the surface of SGR J1830−0645: an extended warm region (average blackbody temperature ∼0.45 keV and radius ∼6 km), and a small hot region (temperature ∼1.1 keV and radius ∼1.5 km). According to our analysis, the modulation of the light curve is (almost) entirely due to the change of the visible area of these regions, the temperatures being fairly constant along the spin phase (Figure 2). The phase-alignment between the light-curve profiles associated with the two blackbody components is indicative of a scenario in which the hot and the warm regions are not spatially separated. A more plausible picture is that what we actually see is a single heated spot with a complex shape where an extended warm region is surrounded by a smaller hotter region with an asymmetric structure (the former producing the relatively broad peak observed in the light-curve profile, the latter resulting in the two prominent narrow peaks separated by ∼0.2 rotational phase cycles; see Figure 2). Indeed, recent 3D simulations have shown that heat injected in a localized region of the crust of a magnetar flows anisotropically to the surface, leading to the appearance of a hot spot with a complex shape and a nonuniform temperature distribution (De Grandis et al. 2020). Soft, thermal photons coming from such a spot can produce a complex pulse profile with a high pulsed fraction, which would be smaller (≲20%) in the case of two circular, antipodal hot spots (Albano et al. 2010; Turolla & Nobili 2013). The pulsed fraction decrease going from lower to higher energies (up to ∼25 keV) is different from what is observed in other magnetars (see, e.g., An et al. 2015, and references therein). Different trends of the pulsed fraction variability with energy can be envisaged within the RCS scenario, depending on the viewing angles as well as the location and velocity distribution of the charged particles in the magnetosphere that up-scatter thermal photons. In the case of SGR J1830−0645, the RCS mechanism tends to wash out the imprint of the (pulsed) primary emission.

Our nondetection of radio pulsations from SGR J1830−0645 is not that surprising per se. Similarly to canonical radio pulsars, radio-loud magnetars are generally characterized by high spin-down luminosities ${\dot{E}}_{\mathrm{rot}}\gtrsim {10}^{33}$ erg s⁻¹, implying ${L}_{{\rm{X}},{\rm{q}}}/{\dot{E}}_{\mathrm{rot}}\lt 1$ (Rea et al. 2012; the only outlier being XTE J1810−197). A reliable estimate of ${L}_{{\rm{X}},{\rm{q}}}/{\dot{E}}_{\mathrm{rot}}$ for SGR J1830−0645 is difficult owing to uncertainties on the distance and the nondetection of the source in pre-outburst X-ray data. However, if its quiescent luminosity is not much below our quoted limit, it would have ${L}_{{\rm{X}},{\rm{q}}}/{\dot{E}}_{\mathrm{rot}}\gg 1$, in line with being radio-silent. Alternatively, SGR J1830−0645 might be radio-loud but undetectable due to unfavorable beaming.

Future high-sensitive X-ray observations will be key in mapping the evolution of the heated spot on the star surface and the long-term contribution of magnetospheric currents to the broadband X-ray emission of SGR J1830−0645.

We thank N. Schartel and F. Harrison for approving Target of Opportunity observations with XMM-Newton and NuSTAR in the Director's Discretionary Time, and the XMM-Newton and NuSTAR SOCs for carrying out the observations. We also thank B. Cenko and the Swift duty scientists and science planners for making the Swift Target of Opportunity observations possible. This research is based on observations with XMM-Newton (ESA/NASA), NuSTAR (CalTech/NASA/JPL), and Swift (NASA/ASI/UKSA) and on data retrieved through the NASA/GSFC's HEASARC archives. The Sardinia Radio Telescope is funded by the Italian MIUR, ASI, and the Autonomous Region of Sardinia, and is operated as a National Facility by INAF. We used data collected at the Parkes radio telescope (proposal No. P1083), part of the Australia Telescope National Facility which is funded by the Australian Government for operation as a National Facility managed by CSIRO. We acknowledge the Wiradjuri people as the traditional owners of the Observatory site. F.C.Z. and A.B. are supported by Juan de la Cierva fellowships. F.C.Z., A.B., N.R., C.D., and D.V. are supported by the ERC Consolidator Grant "MAGNESIA" (No. 817661) and acknowledge funding from grants SGR2017-1383 and PGC2018-095512-BI00. G.L.I., P.E., R.T., and A.T. acknowledge financial support from the Italian MIUR through PRIN grant 2017LJ39LM. G.L.I. also acknowledges funding from ASI-INAF agreements I/037/12/0 and 2017-14-H.O. M.B., A.P., and A.R. acknowledge financial support from the research grant "iPeska" (PI: A. Possenti) funded under the INAF national call PRIN-SKA/CTA with Presidential Decree 70/2016. A.R. acknowledges continuing valuable support from the Max-Planck Society. We acknowledge the support of the PHAROS COST Action (CA16214).

Facility: XMM-Newton (EPIC), Swift (BAT, XRT), NuSTAR, ROSAT (PSPC), Parkes, SRT, ADS, HEASARC.

Software: HEASOFT (v.6.28; Nasa High Energy Astrophysics Science Archive Research Center (Heasarc), 2014), NUSKYBGD (Wik et al. 2014), SAOImageDS9 (v.8.1; Joye & Mandel 2003), SAS (v.19; Gabriel et al. 2004), XSPEC (v.12.11.1; Arnaud 1996), DSPSR (van Straten & Bailes 2011), PSRCHIVE (Hotan et al. 2004), SPANDAK (https://github.com/gajjarv/PulsarSearch), Astropy (v.4.2; Astropy Collaboration et al. 2013, 2018), MATPLOTLIB (v3.3.3; Hunter 2007), NUMPY (v1.19.4; Harris et al. 2020).