The High Time Resolution Universe Survey - XI. Discovery of five recycled pulsars and the optical detectability of survey white dwarf companions

S. D. Bates, D. Thornton, M. Bailes, E. Barr, C. G. Bassa, N. D. R. Bhat,M. Burgay, S. Burke-Spolaor, D. J. Champion, C. M. L. Flynn, A. Jameson,S. Johnston, M. J. Keith, M. Kramer, L. Levin, A. Lyne, S. Milia, C. Ng,E. Petroff, A. Possenti, B. W. Stappers, W. van Straten, C. Tiburzi
Jodrell Bank Centre for Astrophysics, School of Physics and Astronomy, The University of Manchester, Manchester M13 9PL, UK
National Radio Astronomy Observatory, PO Box 2, Green Bank, WV 24944, USA
CSIRO Astronomy & Space Science, Australia Telescope National Facility, P.O. Box 76, Epping, NSW 1710, Australia
Centre for Astrophysics and Supercomputing, Swinburne University of Technology, PO Box 218, Hawthorn, VIC 3122, Australia
ARC Centre of Excellence for All-Sky Astronomy (CAASTRO), Mail H30, Swinburne University of Technology, PO Box 218,
Hawthorn, VIC 3122, Australia
MPI fuer Radioastronomie, Auf dem Huegel 69, D-53121 Bonn, Germany
ASTRON, The Netherlands Institute for Radio Astronomy, Postbus 2, 7990 AA, Dwingeloo, The Netherlands
Iternational Centre for Radio Astronomy Research, Curtin University, Bentley, WA 6102, Australia
INAF - Osservatorio Astronomico di Cagliari, via della Scienza 5, 09047 Selargius, Italy
NASA Jet Propulsion Laboratory, M/S 138-307, Pasadena CA 91106, USA
California Institute of Technology, 1200 E California Blvd., Pasadena, CA, 91125 USA
Department of Physics and Astronomy, West Virginia University, Morgantown, WV, 26506 USA
Dipartimento di Fisica, Università degli Studi di Cagliari, Cittadella Universitaria, 09042 Monserrato (CA), Italy
July 12, 2019

We present the discovery of a further five recycled pulsar systems in the mid-Galactic latitude portion of the High Time Resolution Universe (HTRU) Survey. The pulsars have rotational periods ranging from 2 ms to 66 ms, and four are in binary systems with orbital periods between 10.8 hours and 9.0 days. Three of these binary systems are particularly interesting; PSR J12276208 has a pulse period of 34.5 ms and the highest mass function of all pulsars with near-circular orbits. The circular orbit suggests that the companion is not another neutron star, so future timing experiments may reveal one of the heaviest white dwarfs ever found ( 1.3 M). Timing observations of PSR J14314715 indicate that it is eclipsed by its companion which has a mass indicating it belongs to the redback class of eclipsing millisecond pulsars. PSR J16532054 has a companion with a minimum mass of only , placing it among the class of pulsars with low-mass companions. Unlike the majority of such systems, however, no evidence of eclipses is seen at 1.4 GHz.

pulsars: general - stars: neutron - methods: data analysis
pagerange: The High Time Resolution Universe Survey - XI. Discovery of five recycled pulsars and the optical detectability of survey white dwarf companionsReferencespubyear: 2012

1 Introduction

Although pulsars are commonly born with short spin periods (e.g.  Migliazzo et al., 2002, and references therein), the spin-down rate is such that pulsars are often observed to have spin periods of order . However, the millisecond pulsars (MSPs) spin with periods . The MSPs are located in a region of the - diagram distinct from ordinary pulsars. In order to be rotating so rapidly, MSPs are thought to have undergone a spin-up phase in their evolutionary history. Spin-up involves accretion of matter from an orbiting companion onto the neutron star (NS) (Alpar et al., 1982), which begins when the companion star ages and expands, in some cases, to overflow its Roche lobe. During the accretion phase, systems are thought to be observed as X-ray binaries, both as high-mass (HMXB) and low-mass X-ray binaries (LMXB) depending upon the companion mass.

The evolutionary link between X-ray binaries and MSPs was strengthened with the discovery of SAX J1808.43658, an X-ray binary exhibiting periodic intensity fluctuations with a period of (Wijnands & van der Klis, 1998). More recently, PSR J10230038 was argued to have switched from an LMXB to radio pulsar phase (Archibald et al., 2010), and now has undergone a transition back to an LMXB (Stappers et al., 2014). The radio pulsar J18242425I was the first to be observed to switch to an LMXB phase for a period of around a month, before radio pulsations were once again detected (Papitto et al., 2013). All three systems further strengthen the spin-up model of MSP formation.

Some HMXBs may have a donor sufficiently massive to undergo a core-collapse supernova (ccSN), providing a strong kinematic kick to the system. This kick and mass loss from the companion will either disrupt the binary totally or impart a significant eccentricity to the orbit (Chaurasia & Bailes, 2005). What remains is therefore either an eccentric Double Neutron Star (DNS) system, for example PSR B191316 (Hulse & Taylor, 1975) and PSR J07373039 (Burgay et al., 2003; Lyne et al., 2004), or two separate NSs. In both cases one NS has undergone a spin-up phase (the recycled pulsar) while the other is a young NS (either of which may be observable as a pulsar).

The LMXBs are systems where the companion has a lower mass than the pulsar. The companions reach the end of their lives as a red giant with expulsion of the outer layers of the star, leaving a degenerate core — a white dwarf (WD). The long-lived period (, Tauris & van den Heuvel, 2003) of stable mass transfer acts to circularise the orbit after the ccSN of the NS progenitor (Bhattacharya & van den Heuvel, 1991), explaining why Galactic field MSP-WD binaries usually have a low orbital eccentricity (Phinney, 1992).

The fact that most MSPs are found in binaries is in agreement with the spin-up evolutionary scenario, however of observed MSPs in the Galactic field are isolated. Besides the formation of isolated MSPs from disruption of the orbit during a second ccSN as the HMXB phase ends, MSPs may destroy their companions through ablation due to the wind from the pulsar (Ruderman et al., 1989). Systems undergoing this process are known as black widow systems (Fruchter et al., 1988), consisting of an MSP with an ultra-low-mass companion (). Radio emission from the pulsar is sometimes observed to be eclipsed by ionised material surrounding the companion and, if the orbital inclination is favourable, the companion itself (for example PSR B195720, Fruchter et al. (1988)). The timescale for total ablation of the companion is, however, too long to explain the observed number of isolated MSPs (Eichler & Levinson, 1988). So-called redback binary systems (Roberts, 2011), such as PSR J17405340A in globular cluster NGC6397, also exhibit orbital eclipses of the pulsar emission (D’Amico et al., 2001). redback systems have a more massive companion, () often seen to be non-degenerate in optical studies (Roberts, 2011). These systems have Roche lobe filling factors of -, and as such are possibly an interesting intermediate stage between X-ray binaries and Galactic field MSP binary systems (Breton et al., 2013).

Using the 64-m Parkes radio telescope the HTRU survey (Keith et al., 2010) is providing a comprehensive search of the entire Southern sky with high time and frequency resolution for pulsars. It uses the 20 cm multibeam receiver which was also used by the Parkes Multibeam Pulsar Survey (PMPS) (Manchester et al., 2001) combined with a digital backend system with superior spectral and temporal resolution to the analogue filterbank system used in the PMPS.

This paper outlines the discovery and subsequent observation of five recycled pulsar systems. Section 2 outlines the observations taken to discover and observe these pulsars. In Section 3 we present the results of our timing programme to date, including details of the pulse profiles and, where possible, multi-frequency observations. Finally, in Section 4 we present calculations of predicted WD optical brightnesses for all published MSP-WD systems discovered in the HTRU survey.

Telescope Centre Freq. BW n
(GHz) (GHz) (s)
Parkes 64-m 0.732 0.064 512 900
1.369 0.256 1024 600
3.094 1.024 1024 900
Lovell 76-m 1.524 0.384 768 1800
Table 1: Observing system details for the timing observations made as part of this work; observing bandwidth, BW, number of frequency channels, , and mean observation length, . Note the specifications for the Lovell telescope take into account the removal, as standard, of a section of the observing bandwidth due to contamination by RFI.
Parameter J12276208 J14054656 J14314715 J16532054 J17292117
Right Ascension (J2000) 12:27:00.4413(4) 14:05:21.4255(8) 14:31:44.6177(2) 16:53:31.03(2) 17:29:10.808(6)
Declination (J2000) 62:08:43.789(3) 46:56:02.31(1) 47:15:27.574(4) 20:54:55.1(1) 21:17:28(1)
Galactic longitude () 300.08 315.83 320.05 359.97 4.50
Galactic latitude () 0.59 14.08 12.25 14.26
Discovery signal-to-noise 11.2 125.0 12.0 11.5 13.7
Offset from survey
beam centre (arcmin)
TOA range (MJD) 55901–56641 55668–56557 55756–56627 55658–56676 55505–56686
(ms) 34.52783464780(2) 7.60220343251(1) 2.0119534425332(9) 4.129145284562(2) 66.2928992668(4)
() 18.74(7) 2.79(4) 1.411(3) 1.117(6) 17.2(7)
() 2(1) 0.09(7)
() 18.74(7) 1(1) 1.32(9) 1.117(6) 17.2(7)
DM () 363.0(2) 13.884(3) 59.35(1) 56.56(2) 34.49(4)
DM distance (kpc) 8.32 0.58 1.53 1.63 1.09
(mJy) 0.22 0.92 0.73 0.16 0.20
(mJy kpc) 15.2 0.3 1.7 0.4 0.02
( yr) 5.9 6.1
(erg s)
/ (erg kpc s)
Proper motion:
(mas yr) 44(6) 7(3)
(mas yr) 20(10) 8(4)
(mas yr) 48(8) 11(4)
Binary Model ELL1H ELL1 ELL1 ELL1
Orbital Period (d) 6.721013337(4) 8.95641988(7) 0.4497391377(7) 1.226815259(9)
(lt-s) 23.200663(3) 6.567659(9) 0.550061(2) 0.688855(6)
TASC (MJD) 55991.1937918(2) 55132.23096(2) 55756.1047771(4) 55584.728649(3)
T0 (MJD) 55991.7000(2) 55694.0(4) 55756.23(2) 55584.9(5)
0.0005238(3) 0.000005(2) 0.000023(8) 0.00000(2)
0.0010229(3) 0.000004(2) 0.00000(2)
0.0011494(3) 0.000007(2) 0.000023(8) 0.00001(3)
(deg) 27.11(1) 51(16) 97(18) 0(30)
Shapiro Delay Parameters
Min. () 1.27 0.21 0.12 0.08
Med. () 1.58 0.25 0.14 0.09
RMS of fit (s) 22 26 10 35 162
Reduced 1.30 1.05 1.8 1.7 2.7
When TOAs around superior conjunction are removed
Table 2: Observed and derived parameters for the five MSPs. The DM distance has been estimated using the NE2001 model of Cordes & Lazio (2002), while a pulsar mass of has been assumed in calculating companion masses. Estimates of the contribution to the period derivative from the Shklovskii effect have been included, where significant proper motions have been measured. ELL1 refers to the timing model outlined in the appendix of Lange et al. (2001), and ELL1H is a modification to this model to allow the Shapiro delay to be parameterized in terms of and (see text and Equation 3). Note all parameter errors have been multiplied by .

2 Discovery and Timing

The five pulsars presented here were discovered in the HTRU survey for pulsars and transients (for a full discussion, see Keith et al., 2010). Survey observations were made with the 64-m Parkes radio telescope using the 13-beam multibeam receiver. The survey has an observing band centred at and has a useful bandwidth of . The HTRU survey is split into three areas: the low-, mid- and high-Galactic latitude regions. The discoveries presented here are from the mid-latitude survey, which tiles and with observations of .

The data were processed using the Fourier transform based pulsar search pipeline described in Keith et al. (2010). The discoveries presented here were initially identified using an artificial neural network which aims to highlight the best candidates from the survey for human inspection (Eatough et al., 2010; Bates et al., 2012). After inspection, confirmation observations were taken at the sky positions of the survey beams deemed to contain the best candidates. These confirmations were performed with the Parkes or Lovell telescopes. Four of the five pulsars presented here, PSRs J12276208, J14054656, J14314715, and J16532054, are in orbit with a binary companion, while one, PSR J17292117, is isolated. PSR J12276208 was also independently discovered in the PMPS by two separate groups (Mickaliger et al., 2012; Knispel et al., 2013). All observations, timing, and analysis of PSR J12276208 presented here are independent work.

After confirmation, the new pulsars were timed regularly with Digital Filterbank backend systems; PSRs J16532054 and J17292117 with the 76-m Lovell telescope at Jodrell Bank Observatory (JBO) and PSRs J14314715, J14054656, and J12276208 with the 64-m radio telescope at Parkes (see Table 1). The pulsars timed with the Lovell telescope were observed approximately once per fortnight, whereas those observed using the Parkes radio telescope were observed more sporadically, while maintaining phase coherence, with one case of an approximately 80 day gap between observations.

Each observation resulted in a pulse time-of-arrival (TOA) measurement. Parameters were fitted to the TOAs using the tempo2 software package (Hobbs et al., 2006) and the best fit parameters for the five recycled pulsars are given in Table 2.

3 Results

The pulse periods for the new discoveries range from for PSR J14314715, placing it among the 20 fastest spinning pulsars, to for PSR J17292117, one of the longer periods for a partially recycled pulsar (Manchester et al., 2005) One, PSR J17292117, has not exhibited any detectable periodic variation of the pulse period over the 3 years of timing, indicating that it is one of just two isolated recycled pulsars discovered in the HTRU survey to date. The five pulsars represent some of the different types of known recycled pulsar and binary systems.

Figure 1: A plot of pulse period derivative, , against pulse period, , using the intrinsic values of given in Table 1. Double neutron star (DNS) systems (see Table 3) are plotted as large filled circles, other binaries are small filled circles, and isolated systems are unfilled circles. The discoveries presented in this work are shown as stars, with the isolated system PSR J17292117 an unfilled star. Only non-globular cluster pulsars with and and DNS systems are shown.
Figure 2: Typical pulse profiles are shown for the five millisecond pulsars presented here, from single observations. Profiles have been shifted so that the peaks of the 1400 MHz profile sits at a pulse phase of 0.5. Top, middle, and bottom rows correspond to observing frequencies of , , and , respectively. Blank plots correspond to non-detections. The observing times are given in Table 1. All observations are taken with the 64-m Parkes radio telescope, except for the 1.4 GHz observations and timing of PSRs J16532054 and J17292117, which use the 76-m Lovell telescope at Jodrell Bank.

3.1 Period Derivatives

The measured spin-down rate, , of a pulsar with period can differ from the intrinsic spin-down via the Shklovskii effect (Shklovskii, 1970). A proper motion of , for a pulsar with period at distance , leads to an extra contribution to the period derivative as


where is in units of , which can be a considerable contribution for some MSPs since their rotational periods are so short, the measured so small, and transverse velocities are typically (Toscano et al., 1999). This being the case, the Shklovskii effect is likely to give only small contributions (of the order ) to the measured for PSRs J12276208 and J16532054 (for which no significant proper motion has been measured), but for PSR J17292117, the Shklovskii effect could easily contribute a significant fraction, , of the measured .

In the cases of PSRs J14054656 and J14314715, which have measured, significant, proper motions of 48 and 11 mas yr, respectively. The Shklovskii effect then contributes % and % of the measured for these pulsars. In Table 2 we include corrections for the Shklovskii effect, where it may be calculated. However, the errors on the value of (and, hence, the intrinsic period derivative) are rather large, especially in the case of PSR J14054656, due mainly to contributions from the error on the proper motion and the distance derived from the electron distribution model, which we have assumed to be 30%.

3.2 Psr J12276208

PSR J12276208 is located in the region of the - diagram where mildly recycled pulsars are found (see Figure 1). This is consistent with it having been spun-up during unstable, short-lived mass transfer in an HMXB phase. Pulse profiles from observations at 1.4 and 3.1 GHz are shown in Figure 2.

It has an orbital eccentricity, , orbital period , and a projected semi-major axis, . Assuming , this leads to a minimum companion mass of for an edge-on orbit. The low eccentricity and high minimum companion mass are unusual: there are only three other systems known with and .

PSR J12276208 is similar to other so-called Intermediate Mass Binary Pulsars (IMBPs); for example PSR J14356100 (Camilo et al., 2001) and PSR J22220137 (Boyles et al., 2013). PSR J12276208’s spin parameters are also similar to PSR J06092130 — an isolated pulsar thought to be the result of an HMXB which disrupted during the second ccSN (Lorimer et al., 2004). In an IMBP it is thought the companion was not sufficiently massive to undergo a ccSN, resulting instead in the formation of a heavy CO or ONeMg WD. The minimum mass for the companion to PSR J12276208 is common for ONeMg WDs, which have a mass range around - (Tauris et al., 2012). This is, however, the minimum mass the companion may have; it is already close to the upper limit of ONeMg WDs, and to the Chandrasekhar mass limit.

For inclinations , , and therefore PSR J12276208 might be part of a DNS system. If the companion is a NS, though not one which is visible to date as a pulsar, it would have formed from a massive progenitor (-), and the system would have evolved through an HMXB phase, hence leaving PSR J12276208 partially recycled. Compared to the known DNS systems, listed in Table 3, the eccentricity is small, and the formation of DNS systems with such a small eccentricity appears to be unlikely (Chaurasia & Bailes, 2005). Simulations by Dewi et al. (2005) and Chaurasia & Bailes (2005) do, however, suggest that such low-eccentricity DNS systems could exist because of the possibility of very small and retrograde ccSN kinematic kicks, although they would be extremely rare.

Due to the loss of binding energy (M) during a SN explosion, the mass loss induces a finite eccentricity, , into the system, which is times that of PSR J12276208. The only way this can be lessened is if the exploding star receives a retrograde kick into an exceptionally small volume of phase space to reduce the angular momentum and, hence, eccentricity (Chaurasia & Bailes, 2005). For neutron stars receiving a random kick from a large kick velocity distribution the probability is vanishingly small (0.1%), but the double pulsar’s eccentricity of 0.09 (Lyne et al., 2004) demonstrates that not all neutron stars receive large kicks. Even if the kick was constrained to be small and confined to the pre-SN orbital plane, PSR J12276208’s eccentricity of 0.001 would occur less than of the time as most random kicks act to increase, not decrease, the orbital eccentrity induced by mass loss. On the other hand, such low-eccentricity systems would not lose energy via gravitational wave emission as quickly as eccentric systems, and consequently will survive longer before merging.

(ms) (d) (ls) ()
Table 3: Basic orbital and spin parameters for known DNS systems. Also included are the same parameters for PSR J12276208.

With , the NE2001 electron density model predicts PSR J12276208 to be located at a distance of . This would make an optical detection of a WD companion (for the case of a large orbital inclination) extremely difficult. Consequently, the best chance of measuring the masses in the system comes from the potential measurement of post-Keplerian parameters.

3.2.1 Periastron advance

It is possible to predict relativistic changes to the orbit due to general relativity. One of the easiest post-Keplerian parameters to measure for systems with eccentric orbits is a changing longitude of periastron,


where is the orbital period, is the orbital eccentricity, and and are the companion and pulsar masses respectively. A measured constitutes a measurement of the combined masses, and consequently constrains the companion mass. While the orbital eccentricity of PSR J12276208 is small, it is significantly higher than is typical for systems which have likely evolved through an LMXB phase (Phinney, 1992). Equation 2 predicts for the PSR J12276208 system if we assume pulsar and companion masses of . For a low inclination, and , we would then expect .

Current observations have not made a significant measurement of , however, we can make some quantitative statements about the value of . Using the fake plugin for tempo2, TOAs were generated for PSR J12276208 using the parameters listed in Table 2, varying from 0.01 to 0.05. Fitting those TOAs using tempo2 over the current data span, was only measured at the 2- level for . Since we are unable to make a significant measurement, it is likely that . Extending these simulated TOAs into the future, in all cases another two years of data constrain at the - level, which would allow a direct measurement of and at the 2- level.

3.2.2 Shapiro delay

The radio pulses from a pulsar in a binary system are delayed as they cross the gravitational potential of the companion in a phenomenon called Shapiro delay (Shapiro, 1964). The magnitude of the delay, , is quantified by two post-Keplerian parameters, the range, , and the shape, . For low-eccentricity systems,


where is the orbital phase ( is the ascending node), , , and . The Shapiro delay is the same for every orbit of the system and so when the TOA residuals are folded modulo- a characteristic shape in the timing residuals may be measured.

Since the companion mass for PSR J12276208 is relatively high, implying a large value of , an observing programme was undertaken using the 64-m Parkes radio telescope to obtain TOAs across the expected epoch of superior conjunction, and attempt to observe Shapiro delay in this system. A global timing solution was made (given in Table 2) using the parameterisation for the Shapiro delay given by Freire & Wex (2010), where the fitted parameters and are related to the Shapiro and parameters as


A best-fit value of which is consistent with zero implies a non-detection of the Shapiro delay, and a measurement of allows the orbital inclination to be constrained.

Timing residuals for PSR J12276208, shown in Figure 3 were observed to show some evidence of Shapiro delay due to the apparent non-Gaussianity of the residuals. Fitting for and , we obtained best fit values of and ; however both of these measurements are to a low significance, and corresponding errors in and are large. Nevertheless, the resulting values of range and shape are and .

To estimate when the Shapiro delay might become well-constrained, as in § 3.2.1, TOAs were generated for PSR J12276208 using the Tempo2 plugin fake. Keeping the RMS fixed at 22 s, a further 9 years of timing data were needed in order to constrain both and at the 3- level. Alternatively, using the current data span, and were constrained at the 3- level when the RMS was reduced to 10 s. Therefore, a combination of improved timing and a longer data span may enable precise measurement of and in under 9 years.

Figure 3: Timing residuals from the best fit global model for PSR J12276208 as a function of orbital phase, without any Shapiro delay contribution taken into account. Although the timing residuals appear to show some structure, fitting for the Shapiro delay provides only marginal measurements of orbital parameters.

3.3 Psr J14314715

PSR J14314715 has the 14th shortest spin period of all known pulsars, , and a small period derivative, . These values place PSR J14314715 firmly amongst the fully recycled MSPs in the bottom left of the - diagram (see Figure 1).

Observations of PSR J14314715 at multiple frequencies (see Figure 2) reveal significant pulse profile evolution with observing frequency, as well as evidence of pulse delays and eclipses when the pulsar is at superior conjunction. At the pulse profile is double peaked, with a smaller and wider leading component, the profile at is similar although the leading component appears to be weaker. At (observed away from superior conjunction) the trailing component has become the larger of the two. Strong pulse profile evolution has also been measured in other eclipsing systems, such as PSR J22155135 (Hessels et al., 2011). Further observations of PSR J14314715 at a range of frequencies will prove useful in both further study of the pulse profile evolution and measuring the spectral index.

This pulsar also exhibits significant orbital-phase-dependent delays in the pulse TOAs near superior conjunction (Figure 4). The delays can be explained by an excess DM, attributed to the passage of the radio pulses through ionised material surrounding the companion. The timing model described in Table 2 was generated by excluding TOAs which are obviously associated with the eclipse region, .

Figure 4: The top panel shows the timing residuals for PSR J14314715 which are folded modulo-, between orbital phases of 0 and 0.5. The excess residuals which are not described by the timing model are centred around superior conjunction, an orbital phase of 0.25. The lower panel shows the same delays converted to an electron column density within the eclipsing region. In both panels there are observations at three frequencies; 0.732 GHz (triangle), 1.4 GHz (dots), and 3.1 GHz (circles).

Multi-frequency observations indicate that the magnitude of the eclipse delays depends upon observing frequency, with the observations showing somewhat shorter delays than at . Since this difference is removed when we convert from time delays to additional electrons along the line of sight, we can conclude that these delays are purely due to additional DM introduced by the material between the pulsar and Earth. The lower frequency observations (centred on ) have not resulted in a positive detection to date across the eclipse region (see Figure 4). Treating the delay as purely dispersive, we can relate the delay time, , to the DM as


for observations at a frequency . In Figure 4, this DM has then been converted to the free-electron column density by approximating the depth of the eclipsing region as being equal to the radius of the eclipse region; the computed electron column density is similar to other eclipsing systems; for example PSRs J20510827 (Stappers et al., 1996) and J17311847 (Bates et al., 2011).

Using multiple observations during different eclipses we find that there appear to be significantly different delays at the same orbital phase within the eclipse region. The TOAs were measured during observations which were separated by several months, that is orbital periods. As such, the amount, or density, of dispersive material must be variable on this timescale. The width of the eclipsing region appears to be constant, suggesting it is the free electron density which is variable as opposed to depth of the eclipsing region.

Figure 5: A plot of orbital period against companion mass for non-globular cluster pulsars. Data points correspond to the median mass value, while lower and upper limits correspond to the minimum companion mass and the 90% confidence upper limit. The four binary systems presented in this paper are marked with text, and the other symbols correspond to the systems presented in Table 4. The curves indicate the relationship for HeWDs derived by Tauris & Savonije (1999) for three different companion progenitors, and the vertical lines denote a companion mass of 1.4 M, indicative of a double neutron star system.

From the orbital separation and the mass ratio we determine the approximate distance of the Roche lobe from the companion, , using


from Eggleton (1983), where is the separation of the pulsar and its companion, and is the mass ratio of the system. In this case , about half the size of the eclipse radius. A significant amount of the material in the eclipsing region is therefore outside the companion’s Roche lobe. Material which is close to the Roche lobe can easily be removed by a relativistic pulsar wind (Breton et al., 2013). This would mean some eclipsing material was not gravitationally bound to the companion and must therefore be continually replenished (Stappers et al., 1996).

The spin-down dipole radiation of the pulsar at the distance of the companion is . This is typical of other eclipsing systems in general and in the HTRU sample (Roberts, 2011; Breton et al., 2013). Incident spin down energy at the companion is significantly larger for eclipsing systems than non-eclipsing binaries, an indication that the pulsar’s energy output is at least partly responsible for the dispersive material surrounding the companion.

Name Reference
(kpc) (Gyr)
J1017–7156 2.87 16.69 0.226 0.263 12.55 Keith et al. (2010)
J1056–7117 2.60 6.60 0.150 0.388 12.46 Ng et al. (2014)
J1125–5825 2.58 0.81 0.317 1.765 13.83 Bates et al. (2011)
J1405–4656 0.58 37.64 0.252 0.885 9.70 This paper
J1431–4715 1.53 2.26 0.148 2.290 13.21 This paper
J1431–5740 2.52 10.14 0.187 2.821 14.83 12–16 Burgay et al. (2013)
J1528–3828 2.20 5.00 0.195 1.447 13.16 11–16 Ng et al. (2014)
J1543–5149 2.43 2.02 0.268 4.413 16.34 10–14 Keith et al. (2010)
J1545–4550 2.10 1.08 0.183 2.580 14.19 10–12 Burgay et al. (2013)
J1708–3506 2.77 6.25 0.193 2.534 14.75 11–16 Bates et al. (2011)
J1755–3716 3.90 6.40 0.360 0.400 13.36 11–16 Ng et al. (2014)
J1801–3210 3.95 44.60 0.165 0.687 13.67 Bates et al. (2011)
J1811–2405 1.77 3.15 0.280 11.360 22.60 11–15 Bates et al. (2011)
J1825–0319 3.03 10.61 0.211 9.210 21.62 Burgay et al. (2013)
J2236–5527 0.83 11.40 0.268 0.063 9.66 12–16 Burgay et al. (2013)
Table 4: Predicted optical magnitudes for the HTRU MSPs with helium-core white dwarf companions. Limits are calculated using a lower limit on the pulsar spin-down age of

3.4 Psr J16532054

PSR J16532054 is a fully recycled MSP with and . The pulse profile at 1.4 GHz is shown in Figure 2. The minimum mass of the companion to PSR J16532054 is just , placing it between typical black widow and redback systems. The measurement of is however somewhat lower than for the redback and black widow systems (and PSR J14314715). It is, therefore, possible that the incident pulsar spin-down energy at the surface of the companion is too low to bloat the companion, which would explain why no eclipses have been observed. The orbital period is at least a factor of two longer than known Galactic redback and black widow systems (Roberts, 2011). The wide orbit means that any ionised region would subtend a smaller angle, making eclipses less likely. If, however, the orbit is not being viewed edge-on then the companion mass would be higher, for instance, if then , indicative of a typical HeWD. A detection, or constraint on the magnitude, of the optical companion to this system may help us resolve the nature of the companion and whether this is an example of a wide orbit black widow or redback.

3.5 Psr J17292117

This system, with period and period derivative , has very similar spin parameters to PSR J06092130 (Lorimer et al., 2004) and PSR J22351506 (Camilo et al., 1993). Camilo et al. suggested that PSR J22351506 was spun-up in an HMXB system in which the second ccSN resulted in the disruption of the binary. It may be that similar arguments can be applied to PSR J17292117. As discussed by Burgay et al. (2013), this pulsar was only detected owing to an enhanced flux density due to scintillation during the discovery observation. Subsequent observations have revealed it to have a low luminosity (Table 2).

If this pulsar really was spun up in a binary, then PSR J17292117 might be expected to have a high velocity (Bailes, 1989), and as the system is relatively close (the distance from NE2001 is ), the proper motion should be measurable with further timing. Pulse profiles at 0.7, 1.4 and 3.1 GHz are shown in Figure 2.

3.6 Orbital periods and companion masses

The five MSPs presented here show quite diverse orbital properties. One, PSR 17292117, appears to be isolated and has presumably lost its companion at some point since it was recycled. Three pulsars, PSRs J14054656, J14314715, and J16532054, all have low to intermediate mass companions of to 0.2, and finally PSR J12276208 has a high-mass companion of .

Tauris & Savonije (1999) derived a relationship between orbital period and companion mass for binary MSPs with WD companions which have formed via an LMXB phase (see Figure 5). This relationship is based on a predictable core mass for a main sequence star as a function of stellar radius. During spin-up in an LMXB, the edge of the star is at the position of the Roche lobe, which is a function of only the two masses and the orbital separation (see Equation 6). After the companion star exhausts its fuel and the outer layers are blown off, the stellar core becomes a WD. The WD mass is therefore linked to the orbital period at cessation of mass transfer/spin-up.

Of the binary systems described here, PSR J16532054 has a shorter orbital period than those described by Tauris & Savonije (1999), who only considered . PSRs J14054656 and J14314715 are in agreement with this relationship, and since PSR J12276208 did not form via an LMXB phase, on which this relationship is based, it does not lie on the predicted curve. Indeed, its minimum companion mass is considerably higher than a low-mass WD.

3.7 Non-detection with the Fermi telescope

A large number of MSPs have been discovered by radio observations of point sources first identified by the Large Area Telescope (LAT) on board the Fermi Gamma-Ray Space Telescope (e.g. Ransom et al., 2011; Keith et al., 2011; Cognard et al., 2011). We searched the Fermi LAT Second Source Catalog (Nolan et al., 2012) for point sources which could be associated with the five newly-discovered pulsars presented here, but found no matches. Pulsar detections with Fermi are usually parameterized in terms of (for a spin-down energy loss, , measured in erg s and distance, , in kpc; see  Abdo et al., 2010), which is usually for pulsars that are detected. For PSR J14314715, , so this pulsar might be considered to be right on the margins of detection by Fermi, if indeed the DM distance is correct. The other MSPs presented here all have values of .

4 Possibility of optical detection for HTRU MSP companions

Observations of companions of binary pulsars can allow the determination of parameters which may not be measurable through pulsar timing alone. Of particular interest here are millisecond pulsars with low-mass helium-core white dwarf companions, where a dichotomy in the thickness of the hydrogen envelope surrounding the helium-core of the white dwarf leads to residual hydrogen burning, significantly slowing down the cooling (Alberts et al., 1996; Driebe et al., 1998; Althaus et al., 2001). As a result, helium-core white dwarfs with masses below approximately 0.2 M are typically intrinsically brighter than higher mass white dwarfs. For those systems where the white dwarf is bright enough for optical photometry, it is possible to measure its temperature and cooling age, providing independent constraints on the pulsar age (e.g. van Kerkwijk et al. 2000; Bassa et al. 2003). Furthermore, if the white dwarf has suitable absorption lines, phase-resolved optical spectroscopy of the white dwarf can be used to determine the mass-ratio and model the white dwarf atmosphere to constrain both the white dwarf and pulsar mass (e.g. van Kerkwijk et al. 1996; Callanan et al. 1998; Bassa et al. 2006; Antoniadis et al. 2013, and see van Kerkwijk et al. 2005 for a review).

In Table 4 we present predictions for the apparent brightness of helium-core white dwarf companions to MSPs discovered in the HTRU survey. The binary systems and their properties are listed in Table 4. We use predictions from white dwarf cooling models from Bergeron et al. (1995) and Serenelli et al. (2001) to estimate the absolute magnitude by comparing the predicted white dwarf cooling age with the characteristic spindown age of the pulsar. The masses of the white dwarfs are constrained through the timing measurements of the orbital period and projected semi-major axis. Assuming a pulsar mass of 1.4 M and a range of probable orbital inclinations, we obtain the masses listed in Table 4.

Combining the observed pulsar dispersion measures with the NE2001 model of the Galactic electron distribution (Cordes & Lazio, 2002) yields the distance estimates listed in Table 4. Since some of the HTRU MSPs are at low Galactic latitude, absorption can be significant. The Galactic extinction model of Amôres & Lépine (2005) was used to obtain estimates of the -band absorption for the distance to and line-of-sight of each MSP. Together, these provide the -band distance modulus with which the apparent magnitude can be estimated.

Figure 6 combines all these estimates; the characteristic spindown age of the pulsar with the predicted cooling age of the white dwarf. As pulsar spindown ages may not be a reliable age estimator for the pulsar (Tauris 2012) we conservatively compare a pulsar spin down ages from to . The white dwarf cooling models then predict the absolute -band magnitude . Combining these with the -band distance modulus gives the predicted apparent -band magnitude . These estimates are also listed in Table 4.

Optical detection in imaging observations with an 8 m class telescope typically requires an apparent magnitude less than 24. For white dwarfs brighter than 23rd magnitude optical spectroscopy may be feasible provided suitable absorption lines are present in the spectrum. Based on the estimates from Table 4 and Fig. 5, the HTRU MSPs expected to have the white dwarf companions that can be detected with 8 m class telescopes are PSRs J10567117, J14314715 and J22365527, especially if the characteristic pulsar age is an over-estimate of the real age of the system. In the case of PSR J11255825, a detection may be possible if the characteristic age vastly over-estimates the system’s age.

Figure 6: Comparing characteristic pulsar spindown ages with white dwarf cooling ages. Predictions of white dwarf cooling age as a function of white dwarf mass following the models by Bergeron et al. (1995) (dashed lines/open circles) and Serenelli et al. (2001) (solid lines/closed circles) are shown for a set of absolute -band magnitudes . Plotted with the filled circles are the estimated companion masses (assuming a 1.4 M pulsar and a inclination) against the characteristic spindown age of the pulsar. The vertical line extends from the circles down to to denote a conservative range of pulsar ages. Finally, the size of the circles concentric with the dots denoting the characteristic pulsar age and companion mass, scales with the -band distance modulus , where larger circles indicate lower distance moduli. From the distance modulus and the absolute -band magnitude the apparent -band magnitude for each companion for a given characteristic age can be calculated as . The Hubble time is indicated by the dotted horizontal line at  Gyr.

5 Summary

The discovery of five new MSPs from the High Time Resolution Universe survey has been presented. These pulsars represent members of a wide range of known recycled pulsar types, including possibly the heaviest white dwarf known in orbit around a neutron star, an eclipsing redback system, and an isolated, mildly recycled pulsar, indicating that it probably came from a HMXB which was disrupted during the second ccSN.

Only one of these MSPs, PSR J12276208, is within the PMPS region. This pulsar was in fact co-discovered in recent re-analyses of the PMPS data (Mickaliger et al., 2012; Knispel et al., 2013). The superior temporal and spectral resolution of HTRU data means that even with shorter pointings (in the mid-latitude region) than PMPS, HTRU is able to discover MSPs initially missed by previous surveys which used similar instrumentation to PMPS (Edwards et al., 2001; Jacoby et al., 2006; Burgay et al., 2006).

It would be interesting to attempt detections of some WDs in the HTRU sample in order to test the reliability of MSP spin-down ages. PSRs J10567117, J11255825, J14314715, and J22365527 are the most likely to result in positive detections. An optical observation of PSR J14314715 would also be informative for other reasons; if it is a redback system then its companion may be non-degenerate and it instead could be at the end of the spin-up phase (Roberts, 2011). The companion may also be distinctly non-spheroidal, and measurements of an orbitally modulated light curve can constrain the inclination (Stappers et al., 2001; Reynolds et al., 2007). If spectroscopy as a function of orbital phase is possible for the companion then component masses can be constrained (van Kerkwijk et al., 2011; Romani et al., 2012).


The Parkes Observatory is part of the Australia Telescope which is funded by the Commonwealth of Australia for operation as a National Facility managed by CSIRO. We thank the reviewer, Dipankar Bhattacharya, for suggestions which helped improve the manuscript.


  • Abdo et al. (2010) Abdo A. A. et al., 2010, ApJS, 187, 460
  • Alberts et al. (1996) Alberts F., Savonije G. J., van den Heuvel E. P. J., Pols O. R., 1996, Nature, 380, 676
  • Alpar et al. (1982) Alpar M. A., Cheng A. F., Ruderman M. A., Shaham J., 1982, Nature, 300, 728
  • Althaus et al. (2001) Althaus L. G., Serenelli A. M., Benvenuto O. G., 2001, MNRAS, 324, 617
  • Amôres & Lépine (2005) Amôres E. B., Lépine J. R. D., 2005, AJ, 130, 659
  • Antoniadis et al. (2013) Antoniadis J. et al., 2013, Science, 340, 448
  • Archibald et al. (2010) Archibald A. M., Kaspi V. M., Bogdanov S., Hessels J. W. T., Stairs I. H., Ransom S. M., McLaughlin M. A., 2010, ApJ, 722, 88
  • Bailes (1989) Bailes M., 1989, ApJ, 342, 917
  • Bassa et al. (2006) Bassa C. G., van Kerkwijk M. H., Koester D., Verbunt F., 456, 295
  • Bassa et al. (2003) Bassa C. G., van Kerkwijk M. H., Kulkarni S. R., 403, 1067
  • Bates et al. (2012) Bates S. D. et al., 2012, MNRAS, 427, 1052
  • Bates et al. (2011) Bates S. D. et al., 2011, MNRAS, 416, 2455
  • Bergeron et al. (1995) Bergeron P., Wesemael F., Beauchamp A., 1995, PASP, 107, 1047
  • Bhattacharya & van den Heuvel (1991) Bhattacharya D., van den Heuvel E. P. J., 1991, Phys. Rep., 203, 1
  • Boyles et al. (2013) Boyles J. et al., 2013, ApJ, 763, 80
  • Breton et al. (2013) Breton R. P. et al., 2013, ApJ, 769, 108
  • Burgay et al. (2013) Burgay M. et al., 2013, MNRAS, 433, 259
  • Burgay et al. (2003) Burgay M. et al., 2003, Nature, 426, 531
  • Burgay et al. (2006) Burgay M. et al., 2006, MNRAS, 368, 283
  • Callanan et al. (1998) Callanan P. J., Garnavich P. M., Koester D., 1998, MNRAS, 298, 207
  • Camilo et al. (2001) Camilo F. et al., 2001, ApJ, 548, L187
  • Camilo et al. (1993) Camilo F., Nice D. J., Taylor J. H., 1993, ApJ, 412, L37
  • Chaurasia & Bailes (2005) Chaurasia H. K., Bailes M., 2005, ApJ, 632, 1054
  • Cognard et al. (2011) Cognard I. et al., 2011, ApJ, 732, 47
  • Cordes & Lazio (2002) Cordes J. M., Lazio T. J. W., 2002, ArXiv Astrophysics e-prints (astro-ph/0207156)
  • D’Amico et al. (2001) D’Amico N., Lyne A. G., Manchester R. N., Possenti A., Camilo F., 2001, ApJ, 548, L171
  • Dewi et al. (2005) Dewi J. D. M., Podsiadlowski P., Pols O. R., 2005, MNRAS, 363, L71
  • Driebe et al. (1998) Driebe T., Schoenberner D., Bloecker T., Herwig F., 339, 123
  • Eatough et al. (2010) Eatough R. P., Molkenthin N., Kramer M., Noutsos A., Keith M. J., Stappers B. W., Lyne A. G., 2010, MNRAS, 407, 2443
  • Edwards et al. (2001) Edwards R. T., Bailes M., van Straten W., Britton M. C., 2001, MNRAS, 326, 358
  • Eggleton (1983) Eggleton P. P., 1983, ApJ, 268, 368
  • Eichler & Levinson (1988) Eichler D., Levinson A., 1988, ApJ, 335, L67
  • Freire & Wex (2010) Freire P. C. C., Wex N., 2010, MNRAS, 409, 199
  • Fruchter et al. (1988) Fruchter A. S., Stinebring D. R., Taylor J. H., 1988, Nature, 333, 237
  • Hessels et al. (2011) Hessels J. W. T. et al., 2011, in American Institute of Physics Conference Series, Vol. 1357, Burgay M., D’Amico N., Esposito P., Pellizzoni A., Possenti A., ed, American Institute of Physics Conference Series, p. 40
  • Hobbs et al. (2006) Hobbs G. B., Edwards R. T., Manchester R. N., 2006, MNRAS, 369, 655
  • Hulse & Taylor (1975) Hulse R. A., Taylor J. H., 1975, ApJ, 195, L51
  • Jacoby et al. (2006) Jacoby B. A., Bailes M., Ord S., Knight H., Hotan A., Edwards R. T., Kulkarni S. R., 2006, ApJ, in preparation
  • Keith et al. (2010) Keith M. J. et al., 2010, MNRAS, 1356
  • Keith et al. (2011) Keith M. J. et al., 2011, MNRAS, 414, 1292
  • Knispel et al. (2013) Knispel B. et al., 2013, ApJ, 774, 93
  • Lange et al. (2001) Lange C., Camilo F., Wex N., Kramer M., Backer D. C., Lyne A. G., Doroshenko O., 2001, MNRAS, 326, 274
  • Lorimer et al. (2004) Lorimer D. R. et al., 2004, MNRAS, 347, L21
  • Lyne et al. (2004) Lyne A. G. et al., 2004, Science, 303, 1153
  • Manchester et al. (2005) Manchester R. N., Hobbs G. B., Teoh A., Hobbs M., 2005, AJ, 129, 1993
  • Manchester et al. (2001) Manchester R. N. et al., 2001, MNRAS, 328, 17
  • Mickaliger et al. (2012) Mickaliger M. B. et al., 2012, ApJ, 759, 127
  • Migliazzo et al. (2002) Migliazzo J. M., Gaensler B. M., Backer D. C., Stappers B. W., van der Swaluw E., Strom R. G., 2002, ApJL, 567, L141
  • Ng et al. (2014) Ng C. et al., 2014, MNRAS, 439, 1865
  • Nolan et al. (2012) Nolan P. L. et al., 2012, ApJS, 199, 31
  • Papitto et al. (2013) Papitto A. et al., 2013, The Astronomer’s Telegram, 5069, 1
  • Phinney (1992) Phinney E. S., 1992, Philos. Trans. Roy. Soc. London A, 341, 39
  • Ransom et al. (2011) Ransom S. M. et al., 2011, ApJL, 727, L16
  • Reynolds et al. (2007) Reynolds M. T., Callanan P. J., Fruchter A. S., Torres M. A. P., Beer M. E., Gibbons R. A., 2007, MNRAS, 379, 1117
  • Roberts (2011) Roberts M. S. E., 2011, in American Institute of Physics Conference Series, Vol. 1357, Burgay M., D’Amico N., Esposito P., Pellizzoni A., Possenti A., ed, American Institute of Physics Conference Series, p. 127
  • Romani et al. (2012) Romani R. W., Filippenko A. V., Silverman J. M., Cenko S. B., Greiner J., Rau A., Elliott J., Pletsch H. J., 2012, ApJL, 760, L36
  • Ruderman et al. (1989) Ruderman M., Shaham J., Tavani M., 1989, ApJ, 336, 507
  • Serenelli et al. (2001) Serenelli A. M., Althaus L. G., Rohrmann R. D., Benvenuto O. G., 2001, MNRAS, 325, 607
  • Shapiro (1964) Shapiro I. I., 1964, Phys. Rev. Lett., 13, 789
  • Shklovskii (1970) Shklovskii I. S., 1970, Sov. Astron., 13, 562
  • Stappers et al. (2014) Stappers B. W. et al., 2014, ApJ, 790, 39
  • Stappers et al. (1996) Stappers B. W. et al., 1996, ApJ, 465, L119
  • Stappers et al. (2001) Stappers B. W., van Kerkwijk M. H., Bell J. F., Kulkarni S. R., 2001, ApJ, 548, L183
  • Tauris et al. (2012) Tauris T. M., Langer N., Kramer M., 2012, MNRAS, 425, 1601
  • Tauris & Savonije (1999) Tauris T. M., Savonije G. J., 1999, A&A, 350, 928
  • Tauris & van den Heuvel (2003) Tauris T. M., van den Heuvel E. P. J., 2003, ArXiv Astrophysics e-prints
  • Toscano et al. (1999) Toscano M., Sandhu J. S., Bailes M., Manchester R. N., Britton M. C., Kulkarni S. R., Anderson S. B., Stappers B. W., 1999, MNRAS, 307, 925
  • van Kerkwijk et al. (2005) van Kerkwijk M. H., Bassa C. G., Jacoby B. A., Jonker P. G., 2005, in Astronomical Society of the Pacific Conference Series, Vol. 328, Rasio F. A., Stairs I. H., ed, Binary Radio Pulsars, p. 357
  • van Kerkwijk et al. (2000) van Kerkwijk M. H., Bell J. F., Kaspi V. M., Kulkarni S. R., 2000, ApJ, 530, L37
  • van Kerkwijk et al. (1996) van Kerkwijk M. H., Bergeron P., Kulkarni S. R., 1996, ApJ, 467, L89
  • van Kerkwijk et al. (2011) van Kerkwijk M. H., Breton R. P., Kulkarni S. R., 2011, ApJ, 728, 95
  • Wijnands & van der Klis (1998) Wijnands R., van der Klis M., 1998, Nature, 394, 344
Comments 0
Request Comment
You are adding the first comment!
How to quickly get a good reply:
  • Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
  • Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
  • Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
The feedback must be of minimum 40 characters and the title a minimum of 5 characters
Add comment
Loading ...
This is a comment super asjknd jkasnjk adsnkj
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test description