1. Formation and evolution of magnetars Sergei Popov (SAI MSU) J.A. Pons, J.A. Miralles, P.A. Boldin, B. Posselt, MNRAS (2009) arXiv: 0910.2190 A.Bogomazov, M. Prokhorov Astronomy Reports (2009) arXiv: 0905.3238 MNRAS (2006) astro-ph/0505406

2. Good old classics The pulsar in the Crab nebula A binary system For years two main types of NSs have been discussed: radio pulsars and accreting NSs in close binary systems

3. Diversity of young neutron stars Young isolated neutron stars can appear in many flavors: o Radio pulsars o Compact central X-ray sources in supernova remnants. o Anomalous X-ray pulsars o Soft gamma repeaters o The Magnificent Seven o Unidentified gamma-ray sources o Transient radio sources (RRATs) o Calvera ….

4. Compact central X-ray sources in supernova remnants Cas ARCW 103

5. CCOs in SNRs Age Distance J232327.9+584843 Cas A 0.32 3.3–3.7 J085201.4−461753 G266.1−1.2 1–3 1–2 J082157.5−430017 Pup A 1–3 1.6–3.3 J121000.8−522628 G296.5+10.0 3–20 1.3–3.9 J185238.6+004020 Kes 79 ~9 ~10 J171328.4−394955 G347.3−0.5 ~10 ~6 [Pavlov, Sanwal, Teter: astro-ph/0311526, de Luca: arxiv:0712.2209]arxiv:0712.2209 For two sources there are strong indications for large (>~100 msec) initial spin periods and low magnetic fields: 1E 1207.4-5209 in PKS 1209-51/52 and PSR J1852+0040 in Kesteven 79 [see Halpern et al. arxiv:0705.0978]arxiv:0705.0978

6. Magnetars dE/dt > dE rot /dt dE/dt > dE rot /dt By definition: The energy of the magnetic field is released By definition: The energy of the magnetic field is released Magnetic fields 10 14 –10 15 G

7. Magnetic field estimates Spin down Spin down Long spin periods Long spin periods Energy to support bursts Energy to support bursts Field to confine a fireball (tails) Field to confine a fireball (tails) Duration of spikes (alfven waves) Duration of spikes (alfven waves) Direct measurements of magnetic field (cyclotron lines) Direct measurements of magnetic field (cyclotron lines) Ibrahim et al. 2002

8. Known magnetars SGRs 0526-66 0526-66 1627-41 1627-41 1806-20 1806-20 1900+14 1900+14 0501+4516 – Aug.2008! 0501+4516 – Aug.2008! 1801-23 (?) 1801-23 (?) 0501+4516 (?) 0501+4516 (?) AXPs CXO 010043.1-72 4U 0142+61 1E 1048.1-5937 CXO J1647-45 1 RXS J170849-40 XTE J1810-197 1E 1841-045 AX J1845-0258 1E 2259+586 1E 1547.0-5408 (СТВ 109) Catalogue: http://www.physics.mcgill.ca/~pulsar/magnetar/main.html

9. Extragalactic SGRs [D. Frederiks et al. astro-ph/0609544]astro-ph/0609544 It was suggested long ago (Mazets et al. 1982) that present-day detectors could already detect giant flares from extragalactic magnetars. However, all searches in, for example, BATSE database did not provide god candidates (Lazzati et al. 2006, Popov & Stern 2006, etc.). Finally, recently several good candidates have been proposed by different groups (Mazets et al., Frederiks et al., Golenetskii et al., Ofek et al, Crider...., see arxiv:0712.1502 and references therein, for example).arxiv:0712.1502 Burst from M31

10. Magnificent Seven Name Period, s RX 1856 7.05 7.05 RX 0720 8.39 8.39 RBS 1223 10.31 10.31 RBS 1556 6.88? 6.88? RX 0806 11.37 11.37 RX 0420 3.45 3.45 RBS 1774 9.44 9.44 Radioquiet Close-by Thermal emission Absorption features Long periods

11. RRATs 11 sources detected in the Parkes Multibeam Survey (McLaughlin et al 2006) 11 sources detected in the Parkes Multibeam Survey (McLaughlin et al 2006) Burst duration 2-30 ms, interval 4 min-3 hr Burst duration 2-30 ms, interval 4 min-3 hr Periods in the range 0.4-7 s Periods in the range 0.4-7 s Period derivative measured in 7 sources: B ~ 10 12 -10 14 G, age ~ 0.1-3 Myr Period derivative measured in 7 sources: B ~ 10 12 -10 14 G, age ~ 0.1-3 Myr RRAT J1819-1458 detected in the X-rays, spectrum soft and thermal, kT ~ 120 eV (Reynolds et al 2006) RRAT J1819-1458 detected in the X-rays, spectrum soft and thermal, kT ~ 120 eV (Reynolds et al 2006)

12. M7 and other NSs Evolutionary links of M7 with other NSs are not clear, yet. M7-like NSs can be numerous. They can be descendants of magnetars. Can be related to RRATs. Or, can be a different population.

13. Some reviews on isolated neutron stars NS basics: physics/0503245physics/0503245 astro-ph/0405262 SGRs & AXPs: astro-ph/0406133 arXiv:0804.0250 arXiv:0804.0250 CCOs: astro-ph/0311526 arxiv:0712.2209 arxiv:0712.2209 Quark stars: arxiv:0809.4228arxiv:0809.4228 The Magnificent Seven: astro-ph/0609066 arxiv:0801.1143 RRATs: arXiv:0908.3813 Cooling of NSs: arXiv: 0906.1621 astro-ph/0402143 NS structure arXiv:0705.2708 EoS astro-ph/0612440 arxiv: 0808.1279 NS atmospheres astro-ph/0206025 NS magnetic fields arxiv:0711.3650 arxiv:0802.2227 arxiv:0711.3650arxiv:0802.2227 Read the OVERVIEW in the book by Haensel, Yakovlev, Potekhin

14. Too many NSs??? [Keane, Kramer 2008, arXiv: 0810.1512] It seems, that the total birth rate is larger than the rate of CCSN. e - - capture SN cannot save the situation, as they are <~20%. Note, that the authors do not include CCOs. So, some estimates are wrong, or some sources evolve into another. See also astro-ph/0603258.

15. Transient radiopulsar PSR J1846-0258 P=0.326 sec B=5 10 13 G 0802.1242, 0802.1704 Among all rotation powered PSRs it has the largest Edot. Smallest spindown age (884 yrs). The pulsar increased its luminosity in X-rays. Increase of pulsed X-ray flux. Magnetar-like X-ray bursts (RXTE). Timing noise. See additional info about this pulsar at the web-site http://hera.ph1.uni-koeln.de/~heintzma/SNR/SNR1_IV.htm However, no radio emission detected. Due to beaming?

16. Magnetars, field decay, heating A model based on field-dependent decay of the magnetic moment of NSs can provide an evolutionary link between different populations (Pons et al.). P Pdot PSRs M7 B=const Magnetars Magnetic fields of NSs are expected to decay due to decay of currents which support them Ohm and Hall decay τ Hall depends on B 0 : τ Hall ~ 1/B 0 Fields decay down to ~2 10 13 G Smaller fields decay by a factor ~2

17. Magnetic field decay vs. thermal evolution arxiv:0710.0854 (Aguilera et al.) Magnetic field decay can be an important source of NS heating. Heat is carried by electrons. It is easier to transport heat along field lines. So, poles are hotter. (for light elements envelope the situation can be different).

18. Extensive population synthesis We want to make extensive population synthesis studies using as many approaches as we can to confront theoretical models with different observational data  Log N – Log S for close-by young cooling isolated neutron stars  Log N – Log L distribution for galactic magnetars  P-Pdot distribution for normal radio pulsars We make calculations for seven different fields, which cover the whole range for young objects. To compare our results with observations we use six different models of field distribution.

19. Cooling curves with decay Magnetic field distribution is more important than the mass distribution.

20. Log N – Log S with heating [The code used in Posselt et al. A&A (2008) with modifications] Log N – Log S for 7 different magnetic fields. 1.3 10 12 G 2. 10 13 G 3. 3 10 13 G 4. 10 14 G 5. 3 10 14 G 6. 10 15 G 7. 3 10 15 G Different magnetic field distributions.

21. Fitting Log N – Log S We try to fit the Log N – Log S with log-normal magnetic field distributions, as it is often done for PSRs. We cannot select the best one using only Log N – Log S for close-by cooling NSs. We can select a combination of parameters.

22. Populations and constraints Birthrate of magnetars is uncertain due to discovery of transient sources. Just from “standard” SGR statistics it is just 10%, then, for example, the M7 cannot be aged magnetars with decayed fields, but if there are many transient AXPs and SGRs – then the situation is different. Limits, like the one by Muno et al., on the number of AXPs from a search for periodicity are very important and have to be improved (the task for eROSITA? MAXI?!). Such limits can be also re-scaled to put constraints on the number of the M7-like NSs and the number of isolated accretors with decayed field. [Muno et al. 2007] L x > 3 10 33 erg s -1

23. Log N – Log L for magnetars Magnetic field distributions: with and without magnetars (i.e. different magnetic field distributions are used). 7 values of initial magnetic field, 8 masses of NSs. SNR 1/30 yrs -1. “Without magnetars” means “no NSs with B 0 >10 13 G”. Non-thermal contribution is not taken into account. Justified but total energy losses. Taking into account periods of transient activity in young magnetars improves the picture.

24. Decay parameters and P-Pdot τ Ohm =10 7 yrs τ Hall =10 2 /(B 0 /10 15 G) τ Ohm =10 6 yrs τ Hall =10 3 /(B 0 /10 15 G) τ Ohm =10 5 yrs τ Hall =10 3 /(B 0 /10 15 G) Longer time scale for the Hall field decay is favoured. It is interesting to look at HMXBs to see if it is possible to derive the effect of field decay and convergence.

25. Realistic tracks and observational evidence We plot realistic tracks for NS with masses 1.4 Msolar. Initial fields are: 3 10 12, 10 13, 3 10 13, 10 14, 3 10 14, 10 15, 3 10 15 G Color on the track encodes surface temperature. Tracks start at 10 3 years, and end at ~3 10 6 years. Kaplan & van Kerkwijk arXiv: 0909.5218

26. Population synthesis of PSRs Best model: = 13.25, σ logB0 =0.6, = 0.25 s, σ P0 = 0.1 s

27. Conclusions-1 There are several different populations of neutron stars which must be studied together in one framework Population synthesis calculations are necessary to confront theoretical models with observations We use different approaches to study different populations using the same parameters distribution In the model with magnetic field decay we focused on log-normal distributions of initial magnetic fields We can describe properties of several populations ◊ close-by cooling NSs ◊ magnetars ◊ normal PSRs with the same log-normal magnetic field distribution Best model: = 13.25, σ logB0 =0.6, = 0.25 s, σ P0 = 0.1 s We exclude distributions with >~20% of magnetars Populations with ~10% of magnetars are favoured

28. Origin of magnetars: We present population synthesis calculations of binary systems. Our goal is to estimate the number of neutron stars originated from progenitors with enhanced rotation, as such compact objects can be expected to have large magnetic fields, i.e. they can be magnetars.

29. A question: 5-10 % of NSs are expected to be binary (for moderate and small kicks) 5-10 % of NSs are expected to be binary (for moderate and small kicks) All known magnetars (or candidates) are single objects. All known magnetars (or candidates) are single objects. At the moment from the statistical point of view it is not a miracle, however, it’s time to ask this question. At the moment from the statistical point of view it is not a miracle, however, it’s time to ask this question. Why do all magnetars are isolated? Two possible explanations Large kick velocities Particular evolutionary path

30. Generation of the magnetic field The mechanism of the magnetic field generation is still unknown. Turbulent dynamo α-Ω dynamo (Duncan,Thompson) α 2 dynamo (Bonanno et al.) or their combination In any case, initial rotation of a protoNS is the critical parameter.

31. Strong field via flux conservation There are reasons to suspect that the magnetic fields of magnetars are not due to any kind of dynamo mechanism, but just due to flux conservation: 1.Study of SNRs with magnetars (Vink and Kuiper 2006). If there was a rapidly rotating magnetar then a huge energy release is inevitable. No traces of such energy injections are found. 2.There are few examples of massive stars with field strong enough to produce a magnetars due to flux conservation (Ferrario and Wickramasinghe 2006) Still, these suggestions can be criticized (Spruit arXiv: 0711.3650)

32. What is special about magnetars? Link with massive stars There are reasons to suspect that magnetars are connected to massive stars (astro-ph/0611589). Link to binary stars There is a hypothesis that magnetars are formed in close binary systems (astro-ph/0505406, 0905.3238). The question is still on the list. AXP in Westerlund 1 most probably has a very massive progenitor >40 Msolar.

33. Progenitor mass for a SGR 0910.4859

34. The first calculations We present population synthesis calculations of binary systems using optimistic assumptions about spinning up of stellar cores and further evolution of their rotation rate. MNRAS vol. 367, p. 732 (2006) An optimistic scenario We run the population synthesis of binaries to estimate the fraction of NS progenitors with enhanced rotation. We use the “Scenario Machine” code. Developed in SAI (Moscow) since 1983 by Lipunov, Postnov, Prokhorov et al. (http://xray.sai.msu.ru/~mystery/articles/review/ (http://xray.sai.msu.ru/~mystery/articles/review/ +arXiv: 0704.1387)

35. The model Among all possible evolutionary paths that result in formation of NSs we select those that lead to angular momentum increase of progenitors. Among all possible evolutionary paths that result in formation of NSs we select those that lead to angular momentum increase of progenitors. Coalescence prior to a NS formation. Roche lobe overflow by a primary without a common envelope. Roche lobe overflow by a primary with a common envelope. Roche lobe overflow by a secondary without a common envelope. Roche lobe overflow by a secondary with a common envelope.

36. Results of calculations

37. Results for the optimistic scenario With an inclusion of single stars (with the total number equal to the total number of binaries) the fraction of ``magnetars'‘ is ~8-14%. Most of these NSs are isolated due to coalescences of components prior to NS formation, or due to a system disruption after a SN explosion. The fraction of ``magnetars'' in survived binaries is about 1% or lower. The most numerous companions of ``magnetars'' are BHs.

38. Problems and questions In these calculations we assume that since a star obtained additional angular momentum, then it is effectively transferred to the core, and it doesn’t loose in afterwards. This is too optimistic. There are three processes (Hirschi et al. 2004, 2005) convection, shear diffusion, meridional circulation which result in slowing down the core rotation. Let us consider more conservative scenarios.

39. GRBs and magnetars It is important to remember that a similar problem – necessity of rapid core rotation – is in explanation of GRB progenitors. We hypothesize that a similar channel is operating in binary systems to produce rapidly rotating pre-SN. If then a BH is born – we have a GRB. If a NS – we have a magnetars. The fraction of magnetars among NSs is similar to the fraction of GRBs among BH-forming SNae. Astronomy Reports vol. 53, p. 325 (2009)

40. Model assumptions Here we consider only tidal synchronization on late stages (end of helium burning, or carbon burning). I.e. a core gets additional momentum not long before the collapse. This is possible only in very narrow systems (P orb <~10 days). We used two laws for stellar wind A. Standard wind C. Enhanced wind for massive stars (classification following arXiv: 0704.1387)

41. Different kicks and mass loss (1)isotropic kick, type A wind scenario; (2) isotropic kick, type C wind scenario; (3) Kick along the spin axis, type A wind scenario; (4) Kick along the spin axis, type C wind scenario Single maxwellian distribution We can fit the fraction of isolated magnetars playing with kick distribution. But these assumptions seem to be artificial.

42. Companions Most of companions are -main-sequence stars (49%) and -black holes (46%). The remaining 5% are roughly equally divided among: -white dwarfs (2%), -Wolf–Rayet stars (1%), -stars filling their Roche lobes (0.7%), -helium stars filling their Roche lobes (the BB stage), -hot white dwarfs (0.7%), -neutron stars (0.6%).

43. Conclusions for the pessimistic scenario We made population synthesis of binary stars to explore the evolution and products of stars with enhanced rotation In the optimistic scenario we easily explain the fraction of magnetars an the fact that they are isolated In a more conservative scenario we need large kicks to explain the fact that all known magnetars are isolated Without detailed data about spatial velocities of magnetars it is difficult to make conclusions Still, it is possible that the channel for magnetar formation is the same as for GRB-progenitors formation, most probably in close binary systems

44. Final conclusions It is possible to explain all main types of neutron stars using one smooth (log-gaussian) initial magnetic field distribution in the framework of decaying fields Explaining the origin of magnetars it is necessary to take into account that all known AXP, SGRs, …,etc. are isolated In the studied models of magnetar origin from binary companions with enhanced rotation it is difficult to explain the data without additional assumptions about kick