Towards a self-consistent theory of pulsar magnetospheric HE/VHE emissions

Transcription

1 Towards a self-consistent theory of pulsar magnetospheric HE/VHE emissions HIROTANI, Kouichi ASIAA/TIARA, Taiwan Nagoya University July 26, 2013 Crab nebula: Composite image of X-ray [blue] and optical [red]

2 1 γ-ray Pulsar Observations After 2008, LAT aboard Fermi has detected more than 117 pulsars above 100 MeV. Fermi/LAT point sources (>100 MeV) Geminga Large Area Telescope Fermi γ -ray space telescope Vela Crab 2nd LAT catalog (Abdo+ 2013)

3 1 γ -ray Pulsar Observations Ground-based, Imaging Air Cherenkov Telescopes (IACTs) found pulsed emission above 25 GeV from the Crab pulsar. VERITAS ( > 120 GeV) Aliu+ (2011, Science 334, 69) MAGIC ( GeV) Aleksić+ (2011a,b) P2 OFF P1 P2 Counts per Bin VERITAS > 120 GeV N. Events 2700 MAGIC, GeV Entries T obs = min 2600 Z = (8.6σ) 2 10 H Test = (6.4σ) χ /ndf = /50 (7.7σ) 2400 N ex = Sig = 10.4σ N. Events MAGIC, GeV Entries T obs = min Z = (4.5σ) 2 10 H Test = (4.0σ) 2 χ /ndf = 88.79/50 (3.4σ) N ex = Sig = 6.2σ One NS rotation Counts per Bin Fermi > 100 MeV 1500 Phase Phase One NS rotation MAGIC, GeV Entries T obs = min 2 Z10 = (6.2σ) 1650 H Test = (5.7σ) χ /ndf = /50 (5.1σ) N ex = Sig = 8.3σ 1550 N. Events Phase

4 Pulsed broad-band spectra High-energy (~ GeV) photons are emitted mainly via curvature process by ultra-relativistic, primary e - s/e + s. (created in the particle accelerator) νf ν Crab B Vela B B yrs NS age (Thompson, EGRET spectra) Geminga B yrs ev hν 100 MeV

5 Pulsed broad-band spectra High-energy (> 100MeV) photons are emitted mainly via curvature process by ultra-relativistic e ± s. Crab B Vela 10 3 yrs What is the curvature process? B Consider relativistic charges moving along curved B. They emit curvature (Thompson, radiation, EGRET provided spectra) P» P B A charge e with Lorentz factor γ emits the following synchrotron radiation, Geminga 3 c 3 3 c characteristic energy ħω = ħγ yrs radiation power : P : ħω νf ν curv = 2 ħγ curv 3 2c Rc R c B e c = γ 2 ev hν P c 2 r g NS age cf. synchrotron case 3e c = c γ synch 100 MeV3 2 rg 2

6 Pulsed broad-band spectra High-energy (> 100MeV) photons are emitted mainly via curvature process by ultra-relativistic, primary e ± s. However, > 20 GeV, ICS by secondary & tertiary e ± s contributes. Crab B Vela B For pulsar VHE emissions, Klein-Nishina effect becomes important, because ε * i» m e c 2. B yrs NS age Fig. Two Lorentz frames when a photon is up-scattered by a relativisitic e -. Geminga B

7 2 Pulsar Emission Models Let us begin with considering how and where such incoherent, high-energy photons are emitted from pulsars.

8 2 Pulsar Emission Models High-energy emissions are realized when the rotational energy of the NS is electrodynamically extracted and partly dissipated in the magnetosphere. (e.g., unipolar inductor) Magnetic and rotation axes are generally misaligned. Pulsars: rapidly rotating, highly magnetized NS

9 Pulsar emission takes place at Polar gap 2 Pulsar Emission Models (r <30 km), near NS surface Outer gap, or slot gap (r ~10 3 km), near the light cylinder (outside the null surface) Crab Wind region We neglect the emission from the wind region, because they are not pulsed. Chandra HST

10 Pulsar emission takes place at 2 Pulsar Emission Models Polar gap (r <30 km), near NS surface Outer gap, or slot gap (r ~10 3 km), near the light cylinder (outside the null surface) E arises in a limited volume near the PC surface due to heavy screening by pair discharge. E arises in a greater volume in the higher altitudes due to less efficient pair production.

11 2 Pulsar Emission Models Early 80 s, the polar-cap (PC) model was proposed. (Daugherty & Harding ApJ 252, 337, 1982) A single PC beam can produce a variety of pulse profiles. Open magnetic field lines Rotation axis Radio/ Polar cap beam Outer gap Slot gap Gamma rays

12 2 Pulsar Emission Models Early 80 s, the polar-cap (PC) model was proposed. (Daugherty & Harding ApJ 252, 337, 1982) A single PC beam can produce a variety of pulse profiles. Rotation However, the emission solid angle ( Ω «1 ster) was too axis small to reproduce the wide-separated double peaks. Open magnetic field lines Radio/ Polar cap beam Slot gap Wide-separated double peaks Outer gap Gamma rays

13 2 Pulsar Emission Models Early 80 s, the polar-cap (PC) model was proposed. (Daugherty & Harding ApJ 252, 337, 1982) A single PC beam can produce a variety of pulse profiles. However, the emission solid angle ( Ω «1 ster) was too small to reproduce the wide-separated double peaks. Thus, a high-altitude emission drew attention.

14 2 Pulsar Emission Models To contrive a higher-altitude emission model, the polar gap was extended into higher altitudes (in fact, by hand). Muslimov & Harding (2004a, ApJ 606, 1143; 2004b, ApJ 617, 471) Dyks, Harding & Rudak (2004, ApJ 606, 1125) Harding+ (2008, ApJ ApJ 680, 1378) They explained, e.g., the widely separated double peaks. widely separated double peaks

15 2 Pulsar Emission Models To contrive a higher-altitude emission model, the polar gap was extended into higher altitudes (in fact, by hand). Muslimov & Harding (2004a, ApJ 606, 1143; 2004b, ApJ 617, 471) Dyks, Harding & Rudak (2004, ApJ 606, 1125) Harding+ (2008, ApJ ApJ 680, 1378) They explained, e.g., the widely separated double peaks. However, unfortunately, the higher-altitude SG model contains two fatal electro-dynamical inconsistencies.

16 2 Pulsar Emission Models Problem 1: insufficient luminosity This numerical conclusion is confirmed also analytically (KH 08), showing that a SG can produce only a negligible γ-ray flux. For details, please also refer to my talk last year at CTA-Japan meeting. KH (2008) ApJ 688, L25

17 2 Pulsar Emission Models Problem #2: unphysical assumption of ρ GJ /B In the SG model, they assume ρ GJ /B distribution that contradicts with the Maxwell equation. For details, please also refer to my talk last year at CTA-Japan meeting. Unfortunately, this assumption is unphysical.

18 2 Pulsar Emission Models Problem #2: unphysical assumption of ρ GJ /B In fact, to solve the insufficient flux problem (prob. #1), a geometrically thick version of the higher-altitude SG model, the pair-starved PC (PSPC) model, was proposed (Venter+ 2009, ApJ 707, 800). However, the PSPC model adopts the same ρ GJ /B distribution, which means that the same difficulty applies. Higher-latitude SG model & PSPC model contradict with Maxwell eq. That is, higher-altitude extension of the polar-cap model failed.

19 2 Pulsar Emission Models As an alternative possibility of high-altitude emission model, the outer gap model was proposed. Cheng, Ho, Ruderman (1986, ApJ 300, 500) So far, there have been found no serious electro-dynamical problems in the OG model (unlike SG or PSPC model). Thus, let us concentrate on the OG model in what follows.

20 Indeed, the sub-tev components from the Crab pulsar shows that pulsed γ-rays are emitted from the outer magnetosphere (γb ee). We thus consider the outer-gap model (Cheng+ 86, ApJ 300,500) in this talk. 2 1 ] 2 Pulsar Emission Models dn/de [ TeV cm E s this work this work Crab (P1+P2) (a) Crab Pulsar, P1+P2 (P1+P2) M MAGIC Stereo, this work (P1+P2) E MAGIC Stereo, this work MAGIC-mono Fermi-LAT Fermi-LAT VERITAS (P1+P2) E MAGIC Mono (Aleksić et al. 2011) (P1+P2) M Fermi-LAT (Abdo et al. 2010) (P1+P2) E Fermi-LAT (Aleksić et al. 2011) (P1+P2) V VERITAS (Aliu et al. 2011) Nebula, MAGIC Stereo, this work Nebula, MAGIC Stereo, ICRC 2011 Nebula, Fermi-LAT (Abdo et al. 2010) tot. pulsed, OG+pairs (this work) Aleksic + (2012) A&A 540, A Energy [ GeV ] syst.

21 2 Pulsar Emission Models Various attempts have been made on recent OG model: 3-D geometrical model phase-resolved spectra (Cheng + 00; Tang + 08) atlas of light curves for PC, OG, SG models (Watters + 08) 2-D self-consistent solution (Takata + 06; KH 06) 3-D self-consistent solution phase-resolved spectra, absolute luminosity if we give only P, dp/dt, α, kt (+ζ ) (this talk) In this talk, I ll present the most recent results obtained in my 3-D version of self-consistent OG calculations.

22 2 Pulsar Emission Models 3-D self-consistent OG model Death line of normal and millisecond PSRs on (P,P) plane (Wang & KH 2011, ApJ 736, 127) Spectral hardening of trailing light-curve peak (KH 2011, ApJ 733, L49) Evolution of γ-ray luminosity of rotation-powered PSRs (KH 2013, ApJ 766, 98) Crab pulsars HE-VHE pulsed emission (Alkesic , ApJ 742, 43; Alkesic , AA 540, A69) Comment on Lorentz invariance violation tests. Today s talk I. Today s talk II.

23 3 Modern Outer-gap Model: Formalism Self-sustained pair-production cascade in a rotating NS magnetosphere: e ± s are accelerated by E Relativistic e + /e - emit γ-rays via synchro-curvature, and IC processes γ-rays collide with soft photons/b to materialize as pairs in the accelerator

24 3 Modern OG Model: Formalism ψ ψ ψ ψ 4 π ( ρ ρgj ), x y z = = where E ψ ρ GJ x 2π 1 0 ΩiB,, c [ ] ρ( x) e dγ d χ N ( x, γ, χ) N ( x, γ, χ) + ρ ( x), x Poisson equation for electrostatic potential ψ : = ( x, y, z). π + ion N + /N - : distrib. func. of e + /e - γ : Lorentz factor of e + /e - χ : pitch angle of e + /e -

25 3 Modern OG Model: Formalism Assuming t +Ω φ =0, we solve the e ± s Boltzmann eqs. N v N I ± ± ν + v N ee B S t ± + + = IC + SSC + αν dν dω c p hν together with the radiative transfer equation, di ν dl = I + N ± : positronic/electronic spatial # density, E : mangnetic-field-aligned electric field, S IC : ICS re-distribution function, dω: solid angle element, I ν : specific intensity, l : path length along the ray α ν : absorption coefficient, j ν : emission coefficient j α ν ν ν

26 3 Modern OG Model: Formalism Boundary Conditions To solve the elliptic-type differential eq. (Poisson eq.), we impose Ψ = 0 at inner, lower, upper BDs Ψ = x 0 at outer BD

27 At the inner BD 3 Modern OG Model: Formalism To solve the hyperbolic-type PDE (e ± Boltzmann eq.), we impose N x z γ in + (,, ) = 0 To solve the ODE (radiative transfer eq.), we impose I x z in ν (,, θγ ) = 0, where 0< θ γ < π / 2 (outgoing) That is, no e ± /γ-ray injection across the BD.

28 3 Modern OG Model: Formalism At the outer BD N x z Γ = out (,, ) 0 out ν (,, θγ ) = 0, where π / 2< θ γ < π (in-going) I x z That is, no e ± /γ-ray injection across the BD.

29 4 Gamma-ray vs. Spin-down Luminosities 0.5 First, we demonstrate the observed L L spin. γ L γ L spin nd LAT catalog (Abdo )

30 4 Gamma-ray vs. Spin-down Luminosities To begin with, let us analytically examine the condition for an OG to be self-sustained. An OG emits the energy flux (KH 2008, ApJ 688, L25) µ Ω 1 ( ν Fν ) peak hm, 3 2 c d at distance d by curvature process, where h m denotes dimensionless OG trans-b thickness, µ the dipole moment. OG luminosity can be, therefore, evaluated as µ 2 Ω peak m 3 L 2.36( ν F ) 4π d f 1.23 f h. γ ν Ω Ω c Thus, h m controls the luminosity evolution. E i

31 4 Gamma-ray vs. Spin-down Luminosities To examine h m, consider the condition of self-sustained OG. An inward e - emits N γ in ~10 4 synchro-curvature photons, N γ in τ in ~10 of which materialize as pairs. Each returned, outward e + emits N γ out ~10 5 curvature photons, N γ out τ out ~0.1 of which materialize as pairs.

32 4 Gamma-ray vs. Spin-down Luminosities To examine h m, consider the condition of self-sustained OG. An inward e - emits N γ in ~10 4 synchro-curvature photons, N γ in τ in ~10 of which materialize as pairs. Each returned, outward e + emits N γ out ~10 5 curvature photons, N γ out τ out ~0.1 of which materialize as pairs. That is, gap trans-b-field thickness h m is automatically regulated so that N γ in τ in N γ out τ out =1 is satisfied.

33 4 Gamma-ray vs. Spin-down Luminosities To Step examine 1: Both hn m, in γ consider τ in and Nthe out γ condition τ out are expressed of self-sustained terms OG. of P,µ,α,T, An inward e - and h emits N in m. Thus, γ ~10 4 N in synchro-curvature γ τ in N out γ τ out =1 gives photons, N in γ τ in h ~10 m = h of which m (P,µ,α,T). materialize as pairs. Each returned, outward e + emits N γ out ~10 5 curvature photons, N γ out τ out ~0.1 of which materialize as pairs. That is, gap trans-b-field thickness h m is automatically regulated so that N γ in τ in N γ out τ out =1 is satisfied.

34 4 Gamma-ray vs. Spin-down Luminosities Step 1: express N in γ τ in and N out γ τ out with P,µ,α,T, h m. OG model predicts E µ 2ϖ 3 LC h 2 m Particles (e ± s) saturate at Lorentz factor, 2 3ρ c γ = E 2e. 1/4, emitting curvature photons with characteristic energy, 3 3 hν c = ħc γ. 2 ρ c

35 4 Gamma-ray vs. Spin-down Luminosities Step 1: express N in γ τ in and N out γ τ out with P,µ,α,T, h m. An inward e - or an outward e + emits in out ( ) = 2 / ν, ( ) = 1 / N ee l h N ee l hν γ µ 2 c γ E hm. photons while 2ϖrunning the distance l 2 or l 1. 3 LC c 2 3ρ c γ = E 2e 1/4, hν c 3 3 = ħc γ 2 ρ c.

36 4 Gamma-ray vs. Spin-down Luminosities Step 1: express N γ in τ in and N γ out τ out with P,µ,α,T, h m. An inward e - or an outward e + emits in out ( ) = 2 / ν, ( ) = 1 / N ee l h N ee l hν γ µ 2 c γ E hm. photons while 2ϖrunning the distance l 2 or l 1. 3 LC Such photons materialize as pairs with probability 2 1/4 in 3ρ γ c out τ = l2f = E2σ 2 /, c, τ = l1f1 σ1 / c 2e where F 1, F 2 denotes the X-ray flux and σ 1, σ 2 the pairproduction cross section. Quantities l 1, l 2, F 1, F 2, σ 1, σ 2 can be expressed by P,µ,α,T, and h m, if we specify the B field configuration. c

37 4 Gamma-ray vs. Spin-down Luminosities Step 1: Both N γ in τ in and N γ out τ out are expressed in terms of P,µ,α,T, and h m. Thus, N γ in τ in N γ out τ out =1 gives h m = h m (P,µ,α,T). Step 2: Specifying the spin-down law, P=P(t,α), and the cooling curve, T=T(t), we can solve h m = h m (t, α). Step 3: Independently, P=P(t,α) gives E E( t, α). Step 4: Therefore, we can relate and i i = L = L ( t, ) h E γ γ α i i E = E( t, α) 3 m with intermediate parameter, pulsar age, t. i

38 4 Gamma-ray vs. Spin-down Luminosities Step 2: Give spin-down law and NS cooling curve. Assume dipole-radiation formula, i µ Ω IΩΩ = P = P( t, α) 3 3 c Adopt the minimum cooling scenario (i.e., without any direct-urca, rapid cooling processes). T = T ( t)

39 4 Gamma-ray vs. Spin-down Luminosities Step 2: Cooling curves in the minimum cooling scenario: (contaminated by H, He, C, O) Page +, ApJS 155, 623 (2004)

40 4 Gamma-ray vs. Spin-down Luminosities Step 2: Now we can solve h m = h m (t).

41 4 Gamma-ray vs. Spin-down Luminosities Step 1: Both N γ in τ in and N γ out τ out are expressed in terms of P,µ,α,T, and h m. Thus, N γ in τ in N γ out τ out =1 gives h m = h m (P,µ,α,T). Step 2: Specifying the spin-down law, P=P(t,α), and the cooling curve, T=T(t), we can solve h m = h m (t, α). Step 3: On the other hand, P=P(t,α) gives E E( t, α). Step 4: Therefore, we can relate and L = L ( t, ) h E γ γ α i i E = E( t, α) 3 m i i = with intermediate parameter, pulsar age, t. i

42 4 Gamma-ray vs. Spin-down Luminosities Step 3: We can immediately solve spin-down law. i i = E E( t, α) by the i 2 4 µ Ω E = IΩΩ = C( α) 2 C 3 i i = E 2 = sin α for magnetic-dipole spin-down 2 1+sin for force-free spin-down α c 3 Ω = Ω( t, α) E( t, α)

43 4 Gamma-ray vs. Spin-down Luminosities Step 1: Both N γ in τ in and N γ out τ out are expressed in terms of P,µ,α,T, and h m. Thus, N γ in τ in N γ out τ out =1 gives h m = h m (P,µ,α,T). Step 2: Specifying the spin-down law, P=P(t,α), and the cooling curve, T=T(t), we can solve h m = h m (t, α). Step 3: On the other hand, P=P(t,α) gives E E( t, α). Step 4: Therefore, we can relate and L = L ( t, ) h E γ γ α i i E = E( t, α) 3 m i i = with intermediate parameter, pulsar age, t. i

44 4 Gamma-ray vs. Spin-down Luminosities L h E 3 m Step 4: Use h m =h m (t) to relate with γ i i i E = E( t). Heavy element envelope Light element envelope

45 4 Gamma-ray vs. Spin-down Luminosities Numerical solution is consistent with the analytical one. Lγ ~L spin if L spin >10 36 erg/s However, Lγ declines rapidly if L spin < erg/s KH (2013) ApJ 766, 98

46 4 Gamma-ray vs. Spin-down Luminosities To convert the observed flux F γ into luminosity, L γ = 4π f Ω F γ d 2, it is (conventionally) assumed f Ω =1. However, we underestimate L γ if f Ω >1. View from equator View from rotation axis

47 4 Gamma-ray vs. Spin-down Luminosities To convert the observed flux F γ into luminosity, L γ = 4π f Ω F γ d 2, it is (conventionally) assumed f Ω =1. For example, we obtain f Ω >3 with probability ~50%. ~50% View from equator View from rotation axis

48 4 Gamma-ray vs. Spin-down Luminosities f Ω»1 KH (2013) ApJ 766, 98

49 5 Application to the Crab pulsar We can apply the same numerical scheme to the Crab pulsar. Today, we assume magnetic inclination angle α=60 o, cooling NS surface temperature kt=100ev, (consistent with the cooling curve of a heavy-element envelope)

50 5 Application to the Crab pulsar 3-D OG distribution: trans-b thickness, D., projected on the last-open B field line surface.

51 5 Application to the Crab pulsar Distribution of acceleration E field, E. Max(E ) are projected on the last-open B surface.

52 5 Application to the Crab pulsar Sky map of OG emission Secondary/tertiary synchrotron emission Primary curvature & secondary/tertiary SSC emission Max(E ) are projected on the last-open B surface. Secondary/tertiary SSC emission

53 5 Application to the Crab pulsar If we cut the sky map at a specific viewing angle, we obtain the pulse profile. One NS rotation > 51 GeV > 51 GeV Max(E ) are projected on the last-open B surface. > 51 GeV > 51 GeV

54 5 Application to the Crab pulsar From X-ray observations (of the Crab nebula), ζ~120 o is suggested. Max(E ) are projected on the last-open B surface. Introduce artificial meridional straightening and toroidal bending of B field (due to current): B = (1 c / ) B θ 1 LC θ,vac B = (1 + c / ) B φ ϖ ϖ ϖ ϖ 2 LC φ,vac

55 5 Application to the Crab pulsar Peak separation increases if B field is toroidally bent and meridionally straightened moderately. Max(E ) are projected on the last-open B surface.

56 5 Application to the Crab pulsar Peak separation increases if B field is toroidally bent and meridionally straightened moderately. We can in principle discriminate B field structure near the LC, which has been highly unknown. Max(E ) are projected on the last-open B surface.

57 5 Application to the Crab pulsar Schematic picture of cascading pairs and their emissions:

58 6 Lorentz invariance violation tests Finally, let us consider the Lorentz invariance violation tests using pulsars. Quantum gravity can be tested e.g., by measuring an energy-dependent dispersion relation of mass-less particles. Ex.) Photons would propagate at the speed v( E) c E = 1± E QG For n=1, for instance, two photons with different energies E 1 and E 2 will arrive with time difference, t = L c E E E 2 1 QG n

59 6 Lorentz invariance violation tests L E E L E E t = or E = c E c t QG QG If two photons are emitted at the same place (i.e., same L), we can derive E QG (or set a lower bound of E QG ) from t. Shot-time events (small t) with large photon-energy separation (greater E 2 -E 1 ) are ideal.

60 6 Lorentz invariance violation tests They applied this method to rotation-powered pulsars and examined t at different photon energies, assuming that GeV and TeV photons are emitted from the same location in the magnetosphere. See e.g., Zitzer +, ICRC2013, Rio de Janeiro Otte, ICRC, Beijing McCann +, 4 th Fermi sympo, Monteley Pair production location Max(E ) are projected on the last-open B location surface. Pair production Pair production location

61 6 Lorentz invariance violation tests They applied this method to rotation-powered pulsars and examined t at different photon energies, assuming that GeV and TeV photons are emitted from the same location in the magnetosphere. However, higher energy photons are emitted from higher altitudes by different emission mechanisms. Pair production location Max(E ) are projected on the last-open B location surface. Pair production Pair production location

62 6 Lorentz invariance violation tests Indeed, curvature photons and SSC ones appear close in phase. However, if they appear in different phases, it may merely mean that they are emitted from different locations in the magnetosphere. Max(E ) are projected on the last-open B surface.

63 6 Lorentz invariance violation tests For example, is B field is moderately bent toroidally in the counter-rotation direction, HE and VHE pulses will arrive at different phase. This has nothing to do with the LIV. Max(E ) are projected on the last-open B surface. B = (1 cϖ / ϖ ) B θ 1 LC θ,vac B = (1 + c ϖ / ϖ ) B φ 2 LC φ,vac

64 Summary High-energy pulsar observations have made rapid progress in recent five years by the advent of Fermi/LAT. Development of pulsar emission theory is highly required. Now we can predict the HE emissions from pulsar outer magnetospheres, by solving the set of Maxwell (dive=4πρ) and Boltzmann eqs., if we specify P, dp/dt, α incl, kt NS. The solution coincidentally corresponds to a quantitative extension of classical outer gap model. However, we no longer have to assume the gap geometry, E, e ± distribution functions i i E >10 erg s γ-ray luminosity evolves as Lγ E when, which is consistent with Fermi/LAT observations. Crab pulsar s phase-resolved spectrum can be explained by the current outer-magnetospheric accelerator theory.

65 Thank you.