Student researcher: John Scarpaci '17
Faculty advisor: Dipankar Maitra
Project start/end: May 2014/continuing.
Funding agency: Mars Faculty/Student Research Grant to DM, 2014.
Aim
The main goal of this project is to study time-variation of optical, soft X-ray, and hard X-ray brightnesses Aquila X-1 to better understand the physical origin of its correlated multi-wavelength behavior.Introduction and Background
X-ray binaries (XRBs) are binary star systems that are extremely bright in X-rays when compared to normal stars like the Sun. One of the constituents of an XRB is either a neutron star or a black hole that is actively accreting matter from the other member of the binary system. The other component of an XRB, viz. the donor star, is typically a Sun-like star. As shown in the schematic in Fig.1, matter from the donor spirals onto the accretor (i.e. the neutron star or the black hole) via the formation of an accretion disk. In the very inner parts of the accretion disk, within a few tens of miles of the accretor, the temperature reaches a few million degrees Celsius and emits copious X-rays. The disk is progressively cooler with increasing distance from the center and hence emits lower-energy light in UV, optical, and IR wavelengths.
Fig.1: Schematic of an X-ray binary. Matter from the companion star falls onto the accretor via an accretion disk. The accretor is at the center of the disk. Not all the matter that spirals in falls onto the accretor, but a small fraction sometimes escapes in the form of a jet. (Source: http://chandra.harvard.edu/) |
Fig.2: Multi-wavelength long-term monitoring of Aquila X-1. From top to bottom: J-band, R-band light curves, 1.5–12 keV X-ray brightness measured by RXTE, 2-20 keV brightness from MAXI, and 15-50 keV brightness from Swift. The outburst epochs (periods of high activity) are evident. |
While some XRBs accrete steadily over time, many don't. Quite a few of them spend most of their time in a peaceful slumber (technically known as periods of quiescence), accreting at a very low or non-detectable rates. Once in a while they wake up from their slumber and start accreting vigorously, and brightening up by many orders of magnitude almost all across the electromagnetic spectrum. Such epochs of high activity are known as outbursts.
Our theoretical understanding of the physical mechanisms that drive these outbursts is still quite poor. The widely used Disk Instability Models (DIM) shed some dim light on the ongoing processes, but contrary to their predictions, long-term infrared, optical, and X-ray monitoring of XRBs has revealed that even for one single system the accretion flow varies drastically with time. This is illustrated in Fig. 2 where we show the time-variation of the J-band (IR), R-band (optical), soft X-ray (2-20 keV, from RXTE and MAXI satellites), and hard X-ray (15-50 keV, from Swift satellite) flux from the neutron star XRB system Aquila X-1 over about 9 years. The epochs when the source was actively accreting is evident in all wavelengths.
Data Analysis
[A detailed, weekly log of data analysis was kept by JS, and reviewed by DM. This log is a google document and can be shared upon request.]
X-ray Data: We downloaded the publicly available soft X-ray (i.e. X-rays in the energy range of ~2-10 keV) lightcurves obtained by RXTE/ASM and MAXI/GSC from their respective websites as ASCII text files. For hard X-rays in the range of ~15-50 keV, we used the daily monitoring data from Swift/BAT, also publicly available online. These data sets were then filtered to keep only the good data points with a signal-to-noise ratio of 3 or higher.
Optical Data: For optical data we used the R-band images taken by SMARTS 1.3m telescope at Cerro Tololo Inter-American Observatory (CTIO) in Chile. The SMARTS 1.3m telescope, previously the 2MASS southern telescope, is one of the telescopes operated by the Small and Moderate Aperture Research Telescope System consortium. The SMARTS 1.3m telescope has been used for more than a decade now for nightly monitoring of XRB systems like Aql X-1.
The first step in the optical data analysis was to develop a technique to automatically find Aquila X-1 and two reference stars in our nightly images. This is a nontrivial task because (a) the crowded field-of-view near Aql X-1 contains well over a thousand stars with similar brightnesses, and (b) since the telescope pointing accuracy is about a few arcseconds, the images from one night to another shift randomly by few pixels. Furthermore the SMARTS images do not contain World Coordinate System (WCS; a representation of celestial coordinate system, see these slides by Mark Calabretta for an introduction) information encoded in them, making it difficult to find stars even when their celestial coordinates are well known. Therefore we developed an automated pipeline using software developed and distributed by astrometry.net, to encode WCS information into each image file.
The second step was to determine the brightness of Aql X-1. Once the images were encoded with WCS information, we used the Aperture Photometry Tool to obtain the instrumental brightnesses of Aql X-1 and the two reference stars. Since few nights are photometric (the CTIO defines a photometric night as "one with over 6 continuos hours of cloudless sky"), the brightness of Aql X-1 cannot be determined from the image alone. Instead we used a technique known as differential photometry where we compare the brightness of Aql X-1 with the brightness of nearby reference star(s) to get the true brightness of Aql X-1. This process was also automated so that the instrumental as well as true brightnesses for all the images were computed systematically. The data, codes, and analysis scripts are on the machine astrofs.wheatonma.edu.
The third and final step of the optical data analysis was the subtraction of light from an interloper near Aql X-1. Basically it is known that there exists an interloper star only 0.46" away from the position of the Aql X-1 system. This interloper star is unrelated to the binary and therefore acts like a contaminant in our work. Being a normal star, the interloper's X-ray emission is negligible compared to the X-ray emission from the XRB. However the interloper contributes about 80% of the observed R-band optical light when the XRB is in quiescence. Therefore we subtracted the interloper's flux from our measured brightness to get the true optical brightness of Aql X-1.
Correlating Optical, Soft X-ray, and Hard X-ray Data: The X-ray observations were made using orbiting satellites and are therefore unaffected by weather. Therefore X-ray data is obtained daily throughout the year, with the exception of a few weeks every year when the Sun-satellite-Aql X-1 angle becomes too small for safe observations. The optical observations, in addition to the Sun angle constraint, are also subject to local weather conditions. As the Earth goes around the Sun, the best time to observe a given source also changes. As a result of all these factors mentioned above, the ground-based observations are almost never simultaneous with space-based observations. Based on our previous experience of observing XRBs, we used the following working definition of quasi-simultaneity: if two observations are spaced less than 24 hours apart, then we take these two observations as quasi-simultaneous. Using this definition we analyzed the optical, soft- and hard-X-ray light curves and extracted quasi-simultaneous data. Fig. 3 below is an interactive 3D scatter plot that summarizes our results, where the optical, soft X-rays, and hard X-rays are represented as three spatial dimensions and the red data points correspond to quasi-simultaneous brightnesses in these three different energy ranges.
|
Fig.3:
An interactive 3D plot showing quasi-simultaneous (i.e. obtained less than
24 hours apart) optical, soft X-ray, and hard X-ray brightnesses of Aql X-1
over the past ~8 years. Click and drag mouse to rotate the graph, scroll to
zoom in/out, place mouse on any data point to see the observation date as
well as the brightness values, press 'Esc' to reset to default values. The
menubar near top-right also allows various additional features such as a help menu, saving the current image as PNG, showing data, and additional plot
configuration. |
Results, Ongoing and Future Work
The most striking feature of Fig.3 is the existence of two separate tracks. In a previous work (Maitra & Bailyn 2008), where we had only studied the optical and soft X-ray light curves till 2007, we had noticed hints of these two tracks. But the additional data (now extended to include recent data till 2013), and more importantly the addition of the hard X-ray dimension to the problem (the hard X-ray data comes from Swift mission which became operaional from early 2005), makes the 2-track phenomenon much more obviously noticeable.
We do not yet have a clear explanation of what might be the causing the two tracks that we have found. Based on our current theoretical understanding of XRB systems, the evolution of an XRB during an outburst (and hence the brightnesses at different wavelengths), are driven by a physical mechanism known as disk instability. If only one physical mechanism is at work, then we expect an unique correlation between brightnesses at different wavelengths. The fact that we are observing two separate tracks is directly suggesting that there is not one but two mechanisms.
What could these two mechanisms be? In our previous work we had suggested that some process could be hindering the formation of the very inner part of the accretion disk, and instead channeling much of the accreted matter into an outflow. Recent radio observations have suggested that certain pulsars can switch between two states: an X-ray bright/radio faint state with strong signature of the presence of an inner, X-ray emitting disk, and a radio pulsar phase where the soft X-ray emission is low but the source is detected in radio as a pulsar. While Aql X-1 is not a strong pulsar candidate (although occassional X-ray pulsations have been noticed), some similar switching might be going on in this source.
We are currently in the process of writing our results for publication, re-testing the codes to make sure our observational conclusions are robust, doing a literature survey and discussing these results with colleagues around the world. We expect to submit a paper to a peer-reviewed journal soon. Also we intend JS to present our results in the winter meeting of the American Astronomical Society (to be held in Seattle, WA during the first week of January 2015).
Acknowledgement
This project was funded by a Mars Faculty/Student Research Grant to DM during 2014. In conjunction with a matching grant from the NASA/Rhode Island Space Grant Consortium, the funding also allowed us to procure a 9TB RAID5 linux file server (astrofs.wheatonma.edu). All data and analysis results are backed up on this server. The file server will also of invaluable use for other projects, as well as storing data for classes such as 'Observational Astronomy'. In addition to using data obtained by the CTIO/SMARTS 1.3m telescope, this research has made use of data, software and/or web tools obtained from NASA's High Energy Astrophysics Science Archive Research Center (HEASARC), a service of Goddard Space Flight Center and the Smithsonian Astrophysical Observatory. This research has also made use of the MAXI data provided by RIKEN, JAXA and the MAXI team.