Here we address the questions posed in § 2.3.2:
(1) Are there really as many RRATs in the Galaxy as the initial estimates imply? How well do we know the parameters in Equation 2.37? There does seem to be many pulsars which would be detected as RRATs in surveys such as the PMPS. We now have essentially removed the fRFI factor from Equation 2.37
but, as we have discussed, the beaming and burst rate distributions are also vital ingredients in a population estimate. We note that the population estimate is not to be thought of as representing a distinct group. Interestingly, nulling is not considered when performing population syntheses (basically due to lack of information), even though the discovery of so many nulling sources of late indicates that this may be an important detail to include in evolutionary models. (2) Are they truly a distinct population? What are the implications of this? RRATs are not a distinct population of neutron stars. We have shown that if this were true then the implied Galactic population of neutron stars would be
too large to be accounted for by the supernova rate. The discovery of 19 extra sources in the PMPS also retracts from any idea that the RRATs are a distinct population, but less abundant than previously thought. We propose that RRATs are in fact simply radio pulsars. Although we think of ‘RRAT’ as a detection label, the possibility remains that those pulsars (or a subset thereof), discovered as RRATs, may represent an evolutionary state with a high associated nulling fraction.
(3) Why do they have longer periods than the radio pulsars? Is this signif- icant? When we search for RRATs we make a selection in g − P space which favours the detection of (apparently) nulling pulsars. These searches select high period sources, but we do remain sensitive to short periods (see Figure 9.1). The significance of detecting long period sources may be that it indicates an increased nulling fraction and/or increased pulse-to-pulse modulation for long period/old neutron stars.
(4) What decides whether a NS will manifest itself as a RRAT, as opposed to (say) a magnetar or an XDINS, which occupy the same region of P − ˙P space? We do not yet know the answer to this important question, although the an- swer is fundamental if there is to be “grand unification of neutron stars” (Kaspi, 2010), i.e. the determination of some kind of evolutionary framework. The re- gion of P − ˙P space defined by P = 4 − 10 s, ˙P = 10−13− 10−12 contains radio pulsars (some ‘normal’ pulsars, some RRATs like J1819−1458), magnetars and XDINSs. For very similar spin-down properties we have very different observa- tional manifestations. We might speculate9 that these different classes, although
having similar properties now, have evolved in completely different ways and may have completely different ages. The conditions for coherent radio emission may be very sensitive, with this region a particular area of parameter space on the threshold for emission. This is perhaps consistent with the transient radio emission seen in magnetars and the extreme nulling of the RRATs in this re- gion. If the re-connection rate at the Y-point were slow, or progressed in steps, then bursts of radio emission may be expected between dormant phases, when the magnetospheric configuration was favourable. Regarding the XDINSs, it has been suggested that they may exhibit radio emission but suffer from unfavourable beaming (see e..g. Kondratiev et al. (2008, 2009)), but as their spectra seem to be purely thermal this may not be the case (Haberl, 2007). Only the discovery
of more XDINSs can settle this question convincingly.
(5) Are their observed properties a result of selection effects in our search methods or truly a representation of a class of neutron stars? Given the pa- rameters of our survey and searches, are these the kind of sources we expect to find? We have already discussed the selection in g − P space, but in addition, although there is no selection effect, it seems that sources with high ˙P (and thus high B) are selected. The significance of this is that high-B and/or long-period pulsars are suggested to have either a high null fraction or stronger pulse-to-pulse modulation.
(6) How different is their emission in comparison to the radio pulsar popu- lation? From the currently available data, their radio emission seems to be the same as that seen from pulsars. The only difference seems to be the sporadicity of the emission, implying large nulling fraction and/or pulse-to-pulse modula- tion. The polarisation properties of J1819−1458 are not unusual for its value of
˙
E (Karastergiou et al., 2009). There have not yet been any baseband studies, at the highest time resolution, however the evidence available (see e.g. Chap- ter 7) suggests that their emission is not like that from GRP sources. The X-ray properties are just beginning to be investigated — J1819−1458 has been observed multiple times, but only non-detections have been obtained, so far, for the other 4 sources studied in this band. Very recently, there have also been optical, infrared and γ-ray attempts made, as we have reported in Chapters 5, 7 & 8.
(7) What are there long-term timing properties? How stable, or not, are these? As discussed in Chapter 6, there are now long-term timing solutions for 14 PMPS RRATs and Figure 6.4 shows their distribution in P − ˙P space. For some of the original sources, ∼ 6 years of timing observations are available. For the PMSingle sources this is ∼ 1.5 years. J1819−1458 has shown anomalous glitches, whose significance we are yet to understand, but the other sources have so far shown stable timing properties. In fact some sources, in particular the long- period sources, show remarkably stable timing solutions, especially considering the method of timing via single pulses (see Chapter 5). Monitoring of all sources is ongoing at several telescopes.
(8) Are they old or young? Are they evolutionarily linked to any of the previ- ously known classes of neutron star? There is no single answer for this question regarding age, which covers all of the RRATs. This is perhaps because the ques- tion is suggestive that all of the RRATs are a distinct group or evolutionary stage
when in fact we see a number of ‘solutions’. As we have said, the characteristic ages of the PMPS RRATs are not remarkable in comparison to the overall pul- sar population (see Figure 6.5), but we can comment on the 3 groups that these sources seem to lie in. Bar their sporadicity, the RRATs amongst the normal pulsars seem to have no remarkable properties. The high-B sources are appar- ently young, by the possibly very unreliable measure of characteristic age. The RRATs near the death line are apparently old, by the same measure, something which upcoming X-ray studies of J1840−1419 may shed some further light on.
(9) Can we characterise the observed properties more completely? Are more timing solutions possible and where in P − ˙P space do RRATs really live? Through our monitoring observations over the past 3 years we have been able to characterise the observed radio properties of the known RRATs. In particular, the known timing solutions have increased to 14, up from 3. Chapters 5 & 6 discuss what has been observed, in detail.
(10) Can we discover new sources and improve the characterisations to help to answer all the above questions and identify any key relationships? We have discovered 19 new sources, 12 of which have been observed on multiple occasions as part of a followup campaign of monitoring over the last ∼ 1.5 years. As we have discussed in this chapter, the RRAT searches seem to select highly modulated and/or extreme nulling pulsars. The observed PMPS RRATs can be roughly grouped as: 4 (or perhaps 6) seem to be ‘normal’ pulsars, 4 seem to be high-B radio pulsars (1 of which is above the photon-splitting line) and 4 are old pulsars, some of which are quite close to the death line.
9.3.1
Facts abouts RRATs
We now address a number of assertions, claims and misconceptions concerning the characteristics of RRATs, that we have encountered during the last 3 years, which are held to be correct. Some of these are true, some of these are not.
All RRATs are high-B and therefore linked to magnetars in some way. FALSE. If we arbitrarily define high-B as B ≥ 1013 G, then there are 5 RRATs which
have high magnetic fields. Of those, J1819−1458, with B = 5 × 1013 G, remains the RRAT with the strongest magnetic field strength. Besides the tentative link suggested for J1819−1458, due to its unusual glitches, it is certainly not true to say that links have been identified between the other RRATs and magnetars.
RRATs are only detectable in single pulse searches. FALSE†. Several of the original and PMSingle sources are detectable in periodicity searches, in some cases occasionally and in some cases reliably.
The arrival times of RRAT pulses are random. FALSE†. We have discussed non-random behaviour in the PMSingle sources, where clustering of pulses is seen, e.g. J1724−35 and J1513−5946 (§ 4.5.2). This is also seen in J1913+1330 at Jodrell (§ 5.2.3) and at Parkes (McLaughlin, 2009). Recent work by Palliyaguru et al. (in preparation) has also shown this in 6 of the 8 original RRATs.
Consecutive pulses from RRATs are never seen. FALSE†. One implication of the non-random distribution of pulse amplitudes seen in the Jodrell (see Chap- ter 5) and Parkes (Paliyaguru et al., in preparation, see § 8.1) observations are that we might not see isolated pulses from RRATs. In fact, Palliyaguru et al. report just this — observing higher instances of doublets, triplets and quadru- plets, particularly for J1819−1458, than would be expected by random chance, in the Parkes data for the original RRATs. We can confirm that this is seen in the Jodrell Bank observations of J1819−1458 and J1913+1330 as well as the Parkes observations of the PMSingle sources, as described throughout Chapters 4, 5 & 6. Intriguingly, Palliyaguru et al. also report an instance of detecting pulses from J1819−1458 for 9 consecutive periods. This drastically changes the ‘activation timescales’ needed in some models (although not all, see e.g. Zhang et al., 2007) of RRAT emission, from ∼ 3 ms to ∼ 35 s.
The RRATs are all isolated neutron stars. TRUE. The RRATs discovered so far are all isolated (although see the discussion on the possible origins of J1846−0257 in § 8.1) but this is not in any way a defining feature.
RRATs are special. TRUE. Although this may depend on who you ask.