Example light curves and programme results

Galache et al. (2008) summarise the results of the RXTE monitoring programme be- tween 1999–2008. In total they suggest 14 systems amongst the known population were either newly discovered BeXRB pulsars or were first identified as pulsars by our programme. In addition, they present the long term light curves of every known pulsar at that time which, in many cases, led to a newly determined orbital period from timing analysis of the light curves. In this section, I will briefly summarise the results of the programme between 2008 and the present day and present the light

2_{See http://heasarc.nasa.gov/docs/xte/e-c table.html for the energy conversion table at different}

Figure 2.8: Long-term RXTE/PCA light curve of SXP18.3 in the 3–10 keV band. The panels are described in the text.

curves of some selected systems that show recent X-ray activity. There are 3 main results coming from the programme in the last 3 years:

• the discovery of 4 new pulsars (SXP7.92, SXP11.5, SXP175 & SXP723) and determination of a small number of binary periods of known pulsars.

• the determination of an accurate spin period of the first known pulsar in the Magellanic Bridge.

• the detection and rapid monitoring of Type II outbursts to track spin period evolution in these outbursts for several systems.

The discovery of new pulsars is an important result for more general studies of the SMC population. When a new pulsation is discovered by RXTE, the source needs locating before more detailed studies can be carried out. Chapter 6 discusses what I believe to be the most effective and efficient way of doing this. The latter of these points has proved hugely successful in producing well sampled outbursts to which orbital and spin-up models can be fit. This model fitting to Type II outbursts is the basis on which much of the work presented in this thesis was produced. As I will describe in the following chapters, this adaptation of our traditional observing pro- gramme to target large outbursts has for the first time allowed the determination of

Figure 2.9: Long-term RXTE/PCA light curve of SXP25.5 in the 3–10 keV band. The panels are described in the text.

the orbital parameters of several BeXRB systems in another galaxy. Before present- ing the results of this analysis, I will introduce examples of the data and highlight some specific systems that are currently very active and discuss this in terms of the outburst types seen in BeXRB. The simplistic view of an outburst either being Type I or II as described in Chapter 1, seems to no longer be satisfactory in cate- gorising outbursts in BeXRBs. The light curves presented in Figures 2.8 – 2.12 are used to emphasise some of the more subtle difference in the outbursts we observe. These differences are seen in Galactic systems, but I believe they can only be char- acterised by looking at a large population that have the same age, distance and local environment for a long time. Our SMC dataset offers such an opportunity.

The layout of Figures 2.8 – 2.12 is generic and is produced by the ORCA rou- tine. It is described here so as not to be repeated in each figure caption or the following chapters. The top panel shows the amplitude of the pulsed emission in counts/PCU/s, where each point is a single observation. The vertical blue dashed lines show the most likely X-ray ephemeris based on an orbital period search of the light curve. The second panel shows the source position within the RXTE FOV. A source at the very centre has a collimator of 1 and a source at the very edge 0. The collimator response is approximately linear with distance from the centre. The

Figure 2.10: Long-term RXTE/PCA light curve of SXP46.6 in the 3–10 keV band. The panels are described in the text.

third panel shows the pulse period measurements in seconds. The final panel shows the significance of the detection where the significance, sig, of a given frequency is related to its Lomb-Scargle power, P, by the following formula:

sig= 100 × 1 − e−P
M

(2.3) where M is the number of independent frequencies and is typically 2× 105 _{in our}

analysis pipeline. As such, a LS power& 17 is a detection above 99% significance. Only detections above a significance of 99% are plotted in the third panel for clarity reasons. This 99% limit is denoted in the other panels by red points. Blue points are detections below this threshold. The x-axis is in MJD and spans approximately 14 years. Figures 2.8 – 2.10 are the light curves of SXP18.3, SXP25.5 and SXP46.6 respectively. All three pulsars have been extremely active since their light curves were last published by Galache et al. (2008) and they all show different behaviour. SXP18.3 underwent a massive Type II outburst around MJD54000 (see Schurch et al. 2009 for details) and then turned off. It then switched back on around MJD55000 with a much shorter Type II outburst that was over double the luminosity of the first. This Type II did not end abruptly as before, but reduced into a long series of Type I

Figure 2.11: Long-term RXTE/PCA light curve of SXP59.0 in the 3–10 keV band. The panels are described in the text.

outbursts that peak at around 20% of the Type II luminosity. Timing analysis of this section of the light curve reveals the orbital period of this system at 17.7 d3. This light curve alone is evidence that the traditional Type II outbursts show differences in length and intensity and must be caused by differences in the circumstellar disk at the two epochs. SXP25.5 was rarely detected until MJD∼55000 when it began showing evidence of low luminosity Type I outbursts. These quickly evolved into a more persistent outburst of higher luminosity, though it was still not seen in every observation as a traditional Type II outburst would be. This suggests an intermediate outbursting stage, where a system may be accreting for most of its orbit, but not all of it. This transition also causes the neutron star to strongly spin-up, in contrast to SXP18.3 in which a much more luminous outburst of similar duration has caused almost no spin-up of the neutron star. This implies the torque imparted by the accreted material is different in the two systems. SXP46.6 is the most constant and predictable source in the SMC. Whenever we have had good coverage, we have seen regular Type I outbursts modulated at the binary period. There is a small and constant spin-up and little change in the X-ray luminosity. This is the archetypal Type I outburst as described in the previous chapter.

Figure 2.12: Long-term RXTE/PCA light curve of SXP144 in the 3–10 keV band. The panels are described in the text.

Figures 2.11 – 2.12 show SXP59.0 and SXP144. These systems have shown little X-ray activity since Galache et al. (2008) presented their light curves, but they show more diversity in the type of outburst in their histories. SXP59.0 has under- gone two similar outbursts around MJD51000 and MJD52700. They resemble a persistent (Type II) low luminosity outburst with sharp increases in luminosity at the time of periastron passage (Type I). This shows persistent accretion, but with an increase in accretion rate during periastron. This is very different from SXP18.3 in which the Type II preceded the Type I in a very clear change of state. The spin-up also clearly changes from high to low during the second outburst as the luminosity drops from 5.5 to 2 cts/PCU/s. Finally SXP144 was active for a long time, under- going regular Type I outbursts, similar to SXP46.6. However, in this case there is very definite spin-down of the neutron star whilst accretion is occurring. This is due either to the neutron star spinning in the opposite direction to the accretion flow or the magnetic braking torque of the neutron star is stronger than the torque provided by the accretion flow. Whilst these 5 light curves were chosen as some of the clean- est we have, they demonstrate that care is needed when discussing outburst type. Many factors can affect what type of outburst we see, not least the neutron star spin and its phase in the orbit, the state of the CS disk and the parameters that describe

the binary orbit. Outbursts discussed throughout the rest of this work will often be described in terms of Type I and Type II for ease of understanding, though more specific details will be given when necessary. In the long term however, I believe the nomenclature used in the field should move away from these traditional descrip- tions. In Chapter 7 I will discuss future plans to try and quantify these differences and produce an more thorough classification regime than currently exists.