Simultaneous 2.25/8.60 GHz observations of the newly discovered magnetar - Swift J1818.0$-$1607

Magnetars are a special class of young neutron stars with extremely strong magnetic fields ($10^{13}-10^{15}$ G). They sometimes show dramatic variabilities across the electromagnetic spectrum, especially at high energy band, such as giant flares (Mereghetti, 2008). The decay of enormous internal magnetic fields of magnetars is thought to power these emissions, as they are too bright to be powered by the loss of rotational energy or the accretion power (Duncan & Thompson, 1992). By the end of 2020, there were 31 magnetars and candidates discovered (Olausen & Kaspi, 2014; Coti Zelati et al., 2021)¹¹1http://www.physics.mcgill.ca/~pulsar/magnetar/main.html. Compared with other neutron stars, magnetars generally rotate slower and spin down faster because of their stronger magnetic braking effects (Tong et al., 2013; Gao et al., 2019). Though magnetars show dramatic outbursts at high energy, only 6 magnetars had been detected with radio pulse radiation until the end of 2020 (Olausen & Kaspi, 2014)¹. Typically, magnetar shows variable integrated profiles on timescales from hours to days and relatively flat spectra with typical spectral indices ($\alpha$) greater than $-$0.8 at radio band (Camilo et al., 2007b; Torne et al., 2015, 2017), which is different from normal pulsars that usually show stable integrated profiles and steeper radio spectra with typical $\alpha$ around $-$1.6 (Jankowski et al., 2018).

Swift J1818.0$-$1607 was firstly captured as an X-ray burst by the Swift Burst Alert Telescope on March 12th 2020. It was confirmed to be a magnetar with a spin period ($P$) of 1.36 s by the follow-up observations of Neutron star Interior Composition Explorer (Enoto et al., 2020), making it the magnetar with the shortest $P$ known so far. Furthermore, it was shown to be the 5th radio magnetar known, as its radio pulsations were detected by Effelsberg telescope (Karuppusamy et al., 2020). Shortly after the X-ray burst, its integrated profile could only be detected below 8.0 GHz, though high frequency observations were arranged (Rajwade et al., 2020; Lower et al., 2020b, c; Ridolfi et al., 2020; Majid et al., 2020b; Champion et al., 2020a). About 27 days later, it was detected in several high frequencies observations, such as 8.4 GHz (Majid et al., 2020a), 22 GHz (Liu et al., 2020), 31.9 GHz (Pearlman et al., 2020), 86 and 154 GHz (Torne et al., 2020). Its spectrum got flatter with $\alpha>-0.97$. In most cases, Swift J1818.0$-$1607 showed integrated profile with single peak. Secondary peak component on the right of main peak was sometimes detected at 3.8 GHz (Lower et al., 2020b), 1.37 and 2.55 GHz (Champion et al., 2020a). Furthermore, Lower et al. (2020a) detected integrated profile with triple peaks at 2.368 GHz using Parkes Ultra-Wideband receiver.

As more timing observation data obtained, it was discovered that its rotation frequency derivative $\dot{\nu}$ showed drastic fluctuations, though there was a long-term declining trend of $\nu$. Besides the glitch detected in Swift J1818.0$-$1607 (Champion et al., 2020a; Hu et al., 2020a), it also possibly showed an anti-glitch (Hu et al., 2020a).

Long-term simultaneous multi-frequency observations are vital for studying changes in radio emission of magnetars in both frequency and time. Simultaneous 2.25/8.60 GHz observations of Swift J1818.0$-$1607 were arranged at Shanghai Tian Ma Radio Telescope (TMRT) to do further studies on its radiation and rotational properties. Information about observations is provided in Section 2. The data reduction methods and results are given in Section 3. Finally, further discussions are presented in Section 4.

Simultaneous 2.25/8.60 GHz observations were carried out ranging from MJD 58936 (Mar 28th 2020) to MJD 59092 (Aug 31st 2020) with the TMRT by taking advantage of the parallel working capabilities of the 2.25/8.60 GHz dual-frequency receiver and the digital backend system (DIBAS) which consisted of 3 pairs of analog-to-digital converters and Roach-2 boards (Yan et al., 2018). The dual-frequency receiver is a cryogenically cooled, dual-polarization receiver with the frequency coverage of 2.20-2.30 GHz and 8.20-9.00 GHz, respectively. The total bandwidth was divided into channels with a width of 1 MHz (at 2.25 GHz) and 2 MHz (at 8.60 GHz) with the DIBAS to remove the dispersion effects and radio-frequency interferences (RFIs) (Yan et al., 2015). The incoherent de-dispersion and online-folding observation mode were used in our observations. Each rotation period was divided into 1024 pulse phase bins and folded with 30 s sub-integration length. The observation data were written out as the 8-bit PSRFITS format (Hotan et al., 2004).

After removing RFIs with the pazi command of the Pulsar Archive analysis software (PSRCHIVE) (Hotan et al., 2004), the data was compressed by scrunching all frequency channels and polarizations together with the pam command.

There was no flux calibrator observations arranged before MJD 59039. By assuming the off-pulse root-mean-square (RMS) noise following radiometer equation, we estimated the peak flux density of Swift J1818.0$-$1607 based on typical system equivalent flux density (SEFD), and the SEFD of the TMRT is about 46 Jy and 48 Jy at the corresponding frequencies (Yan et al., 2018). Then, the mean flux density was obtained by dividing the integrated flux density from the on-pulse bins by the total number of phase bins. The uncertainty in mean flux density was obtained from off-pulse RMS noise divided by square root of the number of phase bins. (Zhao et al., 2017). For the observations later than MJD 59039, we estimated its flux density with both SEFD method and calibrator 3C 295. According to our calculations, these two methods are in agreement with each other within the error bar.

Swift J1818.0$-$1607 showed irregular timing properties. As shown in the 2nd panel of Fig. 1, it seems that the $\dot{\nu}$ of Swift J1818.0$-$1607 oscillated but not a stochastic process. The amplitude of $\dot{\nu}$ oscillations seemed to get smaller as time went. Similar timing phenomena were observed in the magnetar 1E 1048.1$-$5937 and 17 ordinary pulsars but with timescales of years (Lyne et al., 2010; Archibald et al., 2020). These $\dot{\nu}$ fluctuations may be caused by certain types of unstable torque that may be affected by stronger magnetic braking (Camilo et al., 2007a; Levin et al., 2012; Tong & Xu, 2013).

Assuming an idealized magnetic dipole model in a vacuum (braking index $n=3$) and much faster spin at birth, its characteristic age $\tau_{\rm c}$ was initially reported as 265, 310 and 240 yr (Champion et al., 2020b; Hu et al., 2020b; Esposito et al., 2020), making it be the youngest magnetar known. The $\dot{\nu}$ of Swift J1818.0$-$1607 fluctuated more than 12 times from MJD 58930 to 58940 (Champion et al., 2020a). Though the $\dot{\nu}$ showed obvious fluctuations in short time-scale, there was obvious linear decrease trend of $\nu$ in long time-scale. Using much longer observation data spanning about 100 days, Hu et al. (2020a) and Champion et al. (2020a) fitted the average $\dot{\nu}$ then got relatively older $\tau_{\rm c}$ age of 470 and 500 yr, respectively. We found the long-term linear decrease of $\nu$ went on in our observations. Base on the best fitted $\dot{\nu}$ obtained with only ours and all the data in Fig. 1, the $\tau_{\rm c}$ inferred with the above assumptions should be 522 and 489 yr, respectively. So, the $\tau_{\rm c}$ of Swift J1818.0$-$1607 is not as young as 300 yr.

Normally, radio-loud magnetars show a flat spectrum typically with $\alpha>-0.8$ (Eatough et al., 2013; Torne et al., 2017). In our observations of Swift J1818.0$-$1607, there was a long-term flux density increase trend at both 2.25 and 8.60 GHz. The $\alpha$ changed from less than $-$1.48 to greater than $-$0.63 as time went on. Judging from plots in Fig. 1, we think its steep spectrum that lasted tens of days was the main reason for no successful detection of integrated profile above 8.0 GHz (Gajjar et al., 2020; Lower & Shannon, 2020; Majid et al., 2020a). The steep radio spectrum was temporarily detected on some magnetars in time-scale of several hours (Lazaridis et al., 2008; Pearlman et al., 2019). A steep magnetar radio spectrum that lasted for tens of days has so far only been observed in J1745$-$2900, soon after its early outburst (Pennucci et al., 2015).

Magnetars usually change their integrated profile shapes dramatically on timescales from hours to days (Pennucci et al., 2015). Besides integrated profile with single-peak and dual-peak structure usually detected at frequencies from 0.8 to 154 GHz, we also detected triple-peaked structure at 2.25 GHz near MJD 59087 as Parkes (Lower et al., 2020a). Though two new components showed at 8.60 GHz, they were not seen at 2.25 GHz from MJD 58936 to 58944 in our observations. Hankins et al. (2015) also detected phenomena in which new component dominated profile at high frequency in the Crab Pulsar.

Lower et al. (2020a) reported two types of mode switching of Swift J1818.0$-$1607 at 0.7-4.0 GHz. Swift J1818.0$-$1607 was quasi-periodically switching between B-mode and Q-mode on MJD 59009, and varying between two longitudinally distinct modes (P- and M-mode) on MJD 59047. Swift J1818.0$-$1607 switched between B-mode and Q-mode synchronously at 2.25/8.60 GHz in 3 days of our observation. Though the mode switching detected by us took place only 3 days later than ‘P-M’ mode switching on MJD 59047, its flux density fluctuation was more similar to ‘B-Q’ mode switching on MJD 59009. But, the right bump of integrated profile appeared in the Q-mode of our observation at 8.60 GHz was not so strong as Lower et al. (2020a)’s result in which it was brighter than the left peak component.

This work was supported in part by the Natural Science Foundation of Shanghai (Grant No. 20ZR1467600), National Natural Science Foundation of China (Grant Nos. U2031119, U1631122 and 11633007), Strategic Priority Research Program of the CAS (XDB23010200) and National Key R&D Program of China (2018YFA0404602).

The data underlying this article will be shared on reasonable request to the corresponding author.