On the morphology of outbursts of accreting millisecond X-ray pulsar Aquila X-1

TL;DR

This paper analyzes the X-ray light curves of two outbursts of the accreting millisecond X-ray pulsar Aquila X-1, revealing that longer quiescent periods lead to more energetic outbursts, with the 2016 event being the most energetic observed.

Contribution

It provides a detailed comparison of two outbursts of Aql X-1, classifies them based on their properties, and investigates the relationship between quiescent duration and outburst energy.

Findings

2016 outburst is the most energetic observed from Aql X-1.

2016 outburst belongs to the long-high class; 2014 to the short-low class.

Longer quiescent periods tend to produce higher peak energy outbursts.

Abstract

We present the X-ray light curves of the last two outbursts --2014 & 2016-- of the well known accreting millisecond X-ray pulsar (AMXP) Aquila X-1 using the monitor of all sky X-ray image (MAXI) observations in the $2-20$ keV band. After calibrating the (MAXI) count rates to the all-sky monitor (ASM) level, we report that the 2016 outburst is the most energetic event of Aql X-1, ever observed from this source. We show that 2016 outburst is a member of the long-high class according to the classification presented by G\"ung\"or et al. with $\sim68$ cnt/s maximum flux and $\sim60$ days duration time and the previous outburst, 2014, belongs to the short-low class with $\sim 25$ cnt/s maximum flux and $\sim 30$ days duration time. In order to understand differences between outbursts, we investigate the possible dependence of the peak intensity to the quiescent duration leading to the outburst and find that the outbursts following longer quiescent episodes tend to reach higher peak energetic.

Full text

††journal: New Astronomy

LMXB low mass X-ray binary AMXP accreting millisecond X-ray pulsar RXTE/PCA proportional counter array RXTE the Rossi X-ray timing explorer MAXI the monitor of all sky X-ray image HMXB high mass X-ray binary SWIFT/XRT the Swift gamma-ray burst mission/X-ray telescope XSPEC an X-ray spectral fitting package FRED fast-rise-exponential-decay DIM the disc instability model LIS low-intensity-state SXT soft X-ray transient HS high/soft state LH low/hard state ASM the all-sky monitor PCU proportional counter unit ISS international space station MSP millisecond pulsar NS neutron star

1 Introduction

Aql X-1, discovered by Kunte et al. [11] in 1973, is a low mass X-ray binary (LMXB) in which a neutron star accretes matter from a disk fed by its K-type companion via Roche lobe overflow [7]. It displays thermonuclear X-ray bursts [10] every few hours due to accumulation of matter on its surface. The burst oscillations [16] as well as the measured spin frequency of $550.273$ Hertz (spin period is $1.8$ msec) [5] indicate to a rapidly spinning neutron star likely spun up in accordance with the recycling hypothesis [1]. The detection of pulsations only during a limited episode indicates the object is an intermittent AMXP.

Aql X-1 is also classified as a soft X-ray transient (SXT) [see 3, for a review] as it shows outbursts almost each year in its X-ray light curve due to the thermal-viscous instability in the accretion disk [see 12, for a review]. We have a wealth of data of these outbursts thanks to the ASM aboard the Rossi X-ray timing explorer (RXTE) which monitors the source since 1996 until the end of the mission, and to the MAXI aboard international space station (ISS) for ongoing observations since 2009. These continuous observations allow Aql X-1 to be a suitable source for studying the outbursts of SXTs.

The morphology of the outbursts of Aql X-1 has been studied by Maitra & Bailyn [13] via optical and near-infrared observations. They identify two types of events; the fast-rise-exponential-decay (FRED) type outbursts that are mostly explained with the disc instability model (DIM) [12, 6] and the low-intensity-state (LIS) events in which the structures of these outburst are in a more complicated variable flux state. which does not exceed the 5 cnt/s level and can last longer than a month.

A broad classification of the FRED type outbursts of Aql X-1 is presented by Güngör et al. [9] (hereafter G14) who showed that FRED type outbursts of Aql X-1 exhibit three main classes depending on the peak flux and the outburst duration: the short-low, the medium-low and the long-high outbursts. The underlying physical cause of the differences between these classes is still unclear.

We present the X-ray light curves of the 2014 and the 2016 outbursts in the light of the classification of G14 in section 2. We explain, in that section, the procedures that we followed to define the outbursts and the durations of the quiescent stages, and the relation between them. We discuss and present the conclusions of our work in section 3.

2 Methodology and Results

We, first, obtained all daily average fluxes from the MAXI [14] and the ASM in the energy range of $1.3-12.1$ keV and $2-20$ keV, respectively. These bands are the largest ranges for each detector. Following G14, we calibrated the MAXI data with the ASM data using the peak count rate of the 2009 and the 2010 outbursts which were observed by both detectors. In Figure 1, we present the long term light curve of Aql X-1 displaying all the outburst of the source since 1996.

We, then smoothed the light curves of the 2014 and the 2016 outbursts with a “natural” spline formalism [15] as done for the earlier outbursts in G14. In Figure 2, we present the set of Aql X-1 outburst morphologies. Correspondingly, based on the outburst classification scheme of G14, we find that the 2014 and the 2016 outbursts fit into the short-low and long-high types, respectively.

We also investigated the possible relation between outburst characteristics and time passed prior to the onset of the activity. The essential step in this investigation was to establish a scheme to define the onset and the end of outbursts. We first selected time intervals with no activity to determine the average count rate and its standard deviation for the quiescent level. Assuming the $3\sigma$ level above quiescence as the threshold to identify a physical change in the light curve, we mark the beginning and the end of outbursts as the first and the last excess above the threshold level, respectively. We also require an outburst episode to have at least five individual measurement for a reliable identification. The brown horizontal lines in Figure 1 indicate the intervals of pre-outburst quiescent stages for FRED type outbursts, and the upside-down triangles mark the onset times of these outbursts.

We find that the maximum intensity of Aql X-1 outbursts is positively correlated with the length of quiescent episode prior to that particular outburst. We present this correlation with square symbols in Figure 3. Quantitatively, we obtain a Spearman’s rank order correlation coefficient of 0.81 with the chance probability of $4.8x10^{-3}$ for the correlation between pre-outburst quiescent duration and the peak flux for FRED type of outbursts. We also fit the peak intensity vs. waiting time trend of these types outbursts with a first order polynomial, which yields a minimum peak rate of 7.9 counts/s and the slope of 0.109$\pm$0.023. Note that such a correlation is not the case for the much lower intensity LIS type nor for FRED+LIS type of outbursts.

3 Discussion & Conclusion

In this study, we updated the classification introduced in G14 by adding the latest two outbursts –2014 & 2016– (Figure 2) of Aql X-1 to the list. We showed that the outburst starting at July 2014 with almost $30$ days duration and $25$ cnt/s maximum flux is a member of the short-low type, and the outburst starting at July 2016 with $60$ days duration and $68$ cnt/s maximum flux is a member of the long-high type. This lends credit to the view that the classification scheme introduced by G14 is robust.

The long-term evolution of the X-ray flux of Aql X-1 shows that the energy released varies from one outburst to another. Although it displays at least one outburst almost each year, we see that the system passed through a relatively quiet episode between 2003 June and 2011 December during when it showed no outburst brighter than $30$ cnt/s (Figure 1), but exhibited many LIS type low energetic events. This propounds that the accretion reservoir is diminishing relatively calmly resulting in low-energetic events rather than leading to FRED type outbursts.

To explore the underlying cause for the differences between outburst types, we searched for a relation between the maximum intensity of the outbursts and the durations of the preceding quiescent episodes. We considered both the FRED and the LIS type events to identify the active and quiescent stages, therefore, we took into account the released energy even via weaker events. Thereby, unlike Campana et al. [4], we introduce possible relation between peak fluxes of the outbursts and the durations of the preceding quiescent episodes as can be seen from Figure 3. The scattered nature of the data in Figure 3 around the linear fit implies that the waiting time is not the only parameter determining the energy to be released in the forthcoming outburst. On the other hand, we observe that the peak fluxes do slightly increase with the durations of the quiescent episodes. In this case, the longer waiting time might lead to accumulation of more material in the disc resulting in a more luminous outburst.

Finally, using the ratio of the peak count rates of the 2016 and 1997 outbursts, and the given peak luminosity of the 1997 outburst in Campana et al. [2], the peak luminosity of the brightest outburst of Aql X-1 in 2016 is estimated as $\sim 8.7\times 10^{37}$erg/s (the distance of the source is 4.5 kpc [8]). This corresponds to a peak luminosity $\sim 0.5$ LEdd for a $1.4$ M⊙ neutron star accretor, after about 512 days of quiescence.

4 Acknowledgement

We thank the anonymous referee for constructive comments. This research has made use of the MAXI data provided by RIKEN, JAXA and the MAXI team and the results provided by the ASM/RXTE teams at MIT and at the RXTE SOF and GOF at NASA’s GSFC.