1: %\documentclaLin_Yi-Fungss[]{article}

2: %\documentstyle[12pt,aasms4]{article}

3: \documentstyle[11pt,aaspp4]{article}

4: %\documentstyle[emulateapj]{article}

5: %\usepackage{emulateapj}

6: \newcommand\beq{\begin{equation}}

7: \newcommand\eeq{\end{equation}}

8: \def\bld#1{\mbox{\boldmath$#1$\unboldmath}}

9: \def\jvec#1{\vec{\bf #1}}

10: \newlabel{fig:correct}{{1}{3}}

11: \newcommand{\figcomment}[1]{#1}

12: %

13: %%

14: %% ** USE THE FOLLOWING DEFINITION FOR FIGURES AND TABLES AT THE END

15: %%

16: %\renewcommand{\figcomment}[1]{}

17: %%

18: %%

19: \def\unsetyr{\def\oyear{\relax}\def\cyear{\relax}\def\cyeara{a\relax}\def\cyearb{b\relax}\def\cyearc{c\relax}\def\cyeard{d\relax}}

20: \def\setyr{\def\oyear{(}\def\cyear{)}\def\cyeara{a)}\def\cyearb{b)}\def\cyearc{c)}\def\cyeard{d)}}

21: \unsetyr

22: \def\jcite#1{\setyr\cite{#1}\unsetyr}

23: 

24: %

25: \def\rmmat#1{{\hbox{\rm #1}}}

26: \def\rmscr#1{\rmmat{\scriptsize #1}}

27: %

28: \newcommand{\be}{\begin{equation}}

29: \newcommand{\ee}{\end{equation}}

30: \newcommand{\bt}{\begin{table} \begin{center}}

31: \newcommand{\et}{\end{center} \end{table}}

32: \newcommand{\ba}{\begin{eqnarray}}

33: \newcommand{\ea}{\end{eqnarray}}

34: \newcommand{\ie}{{\it i.e.~}}

35: \newcommand{\eg}{{\it e.g.~}}

36: \newcommand{\cf}{{\it c.f.~}}

37: %

38: % Math symbols for derivatives

39: \def\mn{\mnras}

40: \def\p{\partial}

41: \def\d{{\rm d}}

42: \def\der#1#2{{\d#1\over\d#2}}

43: \def\dd#1#2{\frac{\d #1}{\d #2}}

44: \def\pp#1#2{\frac{\p #1}{\p #2}}

45: \newcommand{\comment}[1]{\relax}

46: %

47: \def\eqref#1{Equation~\ref{eq:#1}}

48: \def\eqrefb#1{Equation~\ref{#1}}

49: \def\figref#1{Figure~\ref{fig:#1}}

50: \def\figrefb#1{Figure~\ref{#1}}

51: \def\tabref#1{Table~\ref{tab:#1}}

52: \def\bfm#1{\hbox{\boldmath $#1$}}

53: \def\opa{{\tilde \kappa}}

54: \def\rcwx{1E~161348-5055}

55: %

56: 

57: \begin{document}

58: 

59: \title{X-ray emission from middle-aged pulsars}

60: \author{Rosalba Perna$^{1,3}$, Jeremy Heyl$^{2,3}$ \& Lars Hernquist$^3$}

61: \medskip

62: \altaffiltext{1}{Harvard Junior Fellow}

63: \altaffiltext{2}{{\em Chandra} Fellow}

64: \altaffiltext{3}{Harvard-Smithsonian Center for Astrophysics, 60 Garden Street,

65: Cambridge, MA 02138}

66: 

67: 

68: \begin{abstract}

69: We present a simple, unified model which accounts for properties of

70: the X-ray emission from the three middle-aged pulsars PSR 1055-52,

71: PSR 0656+14 and PSR 0630+18 (Geminga).  The X-ray radiation from

72: these objects is pulsed more strongly at energies above a transition point

73: around 0.5~keV.  In addition, the phase of the pulses shifts by about

74: $80^\circ-100^\circ$ degrees around the same point.  Geminga also has the

75: peculiarity that its pulsed fraction {\em decreases} in

76: the 0.3-0.5 keV energy range, attaining a minimum near 0.5~keV.  We

77: show that a two-component hydrogen atmosphere is able to account for

78: these disparate features.  In our model, the hotter component is powered 

79: by particle

80: bombardment and is restricted to the polar regions, while

81: the cooler one covers the entire stellar surface.  The

82: two components also differ in their emission patterns, with the

83: hard and soft contributions coming from areas radiating into

84: fan and pencil beams, respectively.

85: \end{abstract}

86: 

87: \keywords{pulsars: general --- stars: neutron --- X-rays: stars}

88: 

89: \section{Introduction}

90: 

91: X-ray emission from cooling neutron stars (NSs) had been predicted

92: and studied in detail already in the early 1960's (\cite{Chiu64}; 

93: \cite{Tsur64}).

94: However, it was not until the late 70's that the

95: {\em Einstein} Observatory first detected radiation from

96: isolated NSs, and until the {\em ROSAT} era (the 1990's) that more

97: detailed information on the spectral characteristics and pulsed

98: fraction ($Pf$) could be obtained.

99: 

100: Thermal radiation due to cooling can be best observed in a small

101: subset of  pulsars which are

102: middle-aged and have characteristic lifetimes $\tau\sim

103: 10^5$ yr.  In younger pulsars, such as the Crab, X-rays of

104: magnetospheric origin often dominate over thermal emission, while

105: older ones are too cold to have any observable radiation due to

106: cooling. So far, the three best candidates for which

107: thermal emission has likely been detected

108: are PSR 1055-52, PSR 0656+14 and PSR 0630+18.

109: 

110: Apart from a non-thermal, power-law component dominating the

111: high-energy tail (this is particularly evident in PSR 0656+14; \cite{Grei96})

112: and which is thought to be

113: due to magnetospheric processes, the bulk of the emission

114: between $\sim 0.1-1$ keV is well fitted by a double blackbody at two

115: different temperatures.  The ratio of

116: the area of the hotter component to the

117: colder one is typically very small ($\sim$ a few $\times

118: 10^{-5}$ to a few $\times 10^{-3}$).  The colder component has

119: been interpretated as being due to thermal cooling, while the hotter

120: one most likely arises from heated polar caps, with the extra heating being

121: due to bombardment by high-energy particles (\cite{Grei96}).

122: 

123: An important property of the radiation from all three pulsars (for a

124: review see Becker \& Trumper 1997) is a phase shift of $\sim 80^\circ -

125: 100^\circ$ at a transition point which is around 0.5-0.6 keV. The

126: pulsed fraction also shows transitions in its behavior in this

127: region for both PSR 1055-52 and PSR 0656+14. Below $\sim 0.5$ keV, it is

128: roughly constant at the 10\% level, while it rapidly increases after

129: that point to about 20-30\% for PSR 0656+14 and to about 80-90\% for

130: PSR 1055-52. On the other hand, the pulsed fraction from Geminga shows

131: an intriguing feature: in the PSPC channels 8-28 (i.e. roughly at

132: energies below 0.3 keV) the $Pf$ is much larger than in channels 28-53

133: (roughly corresponding to the energy range 0.3-0.5 keV): 33\% versus

134: 20\%.  However, as discussed by \jcite{Page95} and \jcite{Page96},

135: no matter what the surface temperature distribution is, as

136: long as it is not uniform, blackbody emission always gives an increase

137: of the $Pf$ with energy. Detailed modeling by \jcite{Shib95}

138: shows that realistic atmospheres are able to produce a

139: slight decrease of the $Pf$ with increasing energy, but it is still much

140: smaller than what is observed.  A model which could account for the

141: large observed decrease has been presented by \jcite{Page95b}. They

142: assume the presence of warm, magnetized

143: plates on the surface of Geminga, surrounded by cold, 

144: unmagnetized regions, with the warmer

145: plates emitting a softer spectrum than the surrounding areas. In

146: this model, Geminga would therefore have different characteristics

147: than the two other objects in the same class.

148: 

149: In this {\em Letter}, we present a simple model for the emission in

150: the $0.1\sim 1$ keV range that is able to account for the properties

151: of all three pulsars within a unified framework.  We model the cooling

152: component as a blackbody modified by a light-element atmosphere; as

153: discussed in the literature (\cite{Heyl98rxj,Zavl96,Pavl94}),

154: this radiation, which

155: emerges through an atmosphere heated from below, is most likely emitted in

156: a ``pencil'' beam (\cite{Pavl94,Zavl96}).  The hotter component, 

157: due to particle

158: bombardment of the polar caps, is produced in an atmosphere that is hotter

159: on top, i.e. at lower densities. This is due to the fact that

160: the tenuous upper reaches of the neutron-star atmospheres are

161: extremely inefficient emitters; therefore, as the bombarding particles

162: deposit energy in this region, it heats up dramatically to maintain

163: radiative balance.  Meanwhile, the denser layers of the atmosphere

164: which have a larger opacity heat up relatively little.

165: The problem of atmospheres heated from above has been discussed in

166: the literature in the context of slow accretion onto neutron stars

167: (\cite{2000ApJ...537..387Z}) and the X-ray illumination of normal stars in

168: X-ray binaries (\eg\ \cite{1975ApJ...196..583M}).

169: It has been shown that the radiation from an atmosphere illuminated

170: from above may be emitted in a ``fan'' pattern (\eg\ \cite{1975ApJ...196..583M}).

171: In fact, as one moves away from the normal, the effective photosphere of the

172: atmosphere moves to lower densities due to geometrical effects.   In a bombarded

173: atmosphere, these lower densities are hotter, so one sees a hotter and

174: brighter spectrum as one approaches grazing incidence.

175: 

176: We show that a combination of pencil beaming from the cooling

177: component and fan beaming from the hotter polar caps is able to

178: account not only for the magnitude of the pulsations in PSR 1055-52

179: and PSR 0656+14, but also for the phase shift between the soft and the

180: hard components. Moreover, this same model is able to account also for

181: the {\em decrease} of the pulsed fraction observed in the soft X-ray

182: emission of Geminga, without requiring any special model for the

183: composition of the surface of this object.

184: 

185: The paper is organized as follows: in \S2, we describe in detail the

186: X-ray spectral model; specific applications to PSR 1055-52, PSR

187: 0656+14 and Geminga are discussed in \S 3; our results are summarized

188: and discussed in \S 4.

189: 

190: \section{X-ray spectrum}

191: 

192: We model the local emission from the surface of the star, $n(E,T)$,

193: as a blackbody spectrum modified

194: by the presence of an atmosphere, for which we adopt the semianalytical

195: model developed by \jcite{Heyl98atm} with a geometric generalization to

196: study the spectral intensity away from the normal.

197: The dependence of our results on the

198: atmospheric composition will be discussed in \S 3.   

199: We assume that the neutron star has

200: a dipolar magnetic field. With this field geometry and a sufficiently

201: strong intensity, the flux transmitted through the envelope can be

202: approximated as $F\propto\cos^2\psi$, where

203: $\cos^2\psi=4\cos^2\theta_p/(3\cos^2\theta_p+1)$ is the angle between

204: the radial direction and the magnetic field (\cite{Gree83};

205: \cite{Heyl97analns}, 2000); here

206: \beq

207: \theta_p=\arccos(\cos\theta\cos\alpha+\sin\theta\sin\alpha\cos\phi)

208: \label{eq:tetap}

209: \eeq

210: is the angle that the magnetic pole makes with the normal to the

211: star at position $(\theta,\phi)$, while $\alpha$ is the angle that it

212: makes with the line of sight. If $\xi$ is the angle between the 

213: magnetic dipole and the rotation axis, and $\chi$ the angle between

214: the observer's direction and the rotation axis, then the angle $\alpha$

215: is given by 

216: \beq

217: \alpha=\arccos(\cos\chi\cos\xi+\sin\chi\sin\xi\cos\gamma)\;,

218: \label{eq:alpha}

219: \eeq

220: with $\gamma$ being the phase angle.

221: 

222: Outside of the heated polar caps, the local temperature on the star

223: due to thermal cooling is given by 

224: \beq

225: T_{\rm th}(\theta,\phi)=T_p\left[\frac{4\cos^2\theta_p}{3\cos^2\theta_p+1}\;

226: (0.75\;\cos^2\theta_p+0.25)^{0.2}\right]\;,

227: \label{eq:Tdip}

228: \eeq 

229: where, following \jcite{Heyl97analns}, we have assumed a

230: further dependence of the flux on magnetic field strength

231: scaling as $B^{0.4}$. In Equation (3), 

232: $T_p$ is the temperature

233: that the pole would have if there were no reheating. 

234: 

235: Now, let $\beta$ be the angular size of the radius of the polar caps, which are centered

236: around the poles. The hot spot region (in the upper hemisphere)

237: is defined by 

238: \beq

239: \theta\le\beta,\;\;\;\;\;\;\;\;\;\;\;\; \rmmat{if}\;\;\; \alpha=0\;

240: \label{eq:con1}

241: \eeq

242: and

243: \beq

244:    \left\{

245:   \begin{array}{ll}

246:     \alpha-\beta\le\theta\le\alpha+\beta \\

247:       \phi\le\phi_p \;\;\;\rmmat{or}\;\;\;\; 2\pi-\phi\le\phi_p, \;\;\;\;\rmmat{if}

248:         \;\;\;\alpha\ne 0\;\;\;\rmmat{and} \;\;\;\beta\le\alpha\\

249:   \end{array}\right.\;

250: \label{eq:con2}

251: \eeq

252: where

253: \beq

254: \phi_p=\arccos\left[\frac{\cos\beta-\cos\alpha\cos\theta}{\sin\alpha\sin\theta}\right]\;.

255: \label{eq:phip}

256: \eeq

257: Here, we have chosen a system of coordinates such that the prime

258: meridian passes through the North pole of the star, at the center of

259: the upper polar cap.  The hot spot in the lower hemisphere is defined

260: by the condition $\theta\le\pi-\beta$, if $\alpha= 0$, and by

261: Equation (5) and Equation (6) with the

262: substitutions $\alpha \rightarrow \pi-\alpha$ and

263: $\phi\rightarrow\pi-\phi$, if $\alpha\neq 0$. 

264: 

265: Inside the hot spot, we assume the temperature to be constant, $T_\rmscr{hs}(\theta,\phi)

266: ={\rmmat const} \equiv T_{\rm hs}$. As already discussed, 

267: we adopt pencil beaming for the thermal component and fan beaming for the

268: radiation produced in the polar caps, and, for simplicity,  we parameterize them with

269: the functions

270: \beq

271:    \left\{

272:   \begin{array}{ll}

273: f_\rmscr{th}(\delta)\propto \cos^{n_1}(\delta)\\

274: f_\rmscr{hs}(\delta)\propto \sin^{n_2}(\delta)\\

275:   \end{array}\right.\;,

276: \label{eq:beam}

277: \eeq

278: where $\delta$ is the angle that a photon emitted at a colatitude

279: $\theta$ on the star makes with the normal to the surface at the

280: moment of emission. The relation between $\theta$ and $\delta$ 

281: (which is a consequence of the general relativistic effects

282: of light deflection) is

283: given by the ray-tracing function (Page 1995)

284: \beq

285: \theta(\delta)=\int_0^{R_s/2R}x\;du\left/\sqrt{\left(1-\frac{R_s}{R}\right)

286: \left(\frac{R_s}{2R}\right)^2-(1-2u)u^2 x^2}\right.\;,

287: \label{eq:teta}

288: \eeq having defined $x\equiv\sin\delta$.  The value of $n_1$ in the

289: above equation is found to be roughly unity for an unmagnetized

290: light-element atmosphere (\cite{Zavl96}) and slightly larger for a

291: magnetized light-element atmosphere (\cite{Pavl94}), while it is about

292: 0.5 for an atmosphere made of heavy elements (\cite{Raja97}).  In our

293: calculations we restrict ourselves to this range of values.  The

294: exponent $n_2$, on the other hand, is treated as a fit parameter and

295: is allowed to vary from source to source, as one would expect the

296: extent of fan beaming in the bombarded atmosphere to depend on the

297: spatial and energy distribution of the bombarding particles which are

298: unknown (and likely to vary among different objects).

299: 

300: The calculation of the emitted spectrum fully accounts for the

301: consequences of gravitational bending of light

302: and gravitational redshift.  Let us define

303: $e^{-\Lambda_s}=\sqrt{1-R/R_s}$, where $R$ is the radius of the neutron

304: star, and $R_s=2GM/c^2$ is its Schwarzschild radius.  A range of radii

305: compatible with the currently available models for the NS equation of

306: state requires $2\le (R/R_s)\le 4$. Here we take $R$ in this range, with $M=1.4

307: M_\odot$.  Let $D$ be the distance from the star to the observer, and

308: $N_\rmscr{H}$ the intervening column density.  The flux measured by an

309: observer at infinity is then given by the sum of the contributions

310: from the thermal component and from the polar caps

311: \begin{eqnarray}

312: F(E;\gamma)&=&\frac{\pi R^2\;e^{-\Lambda_s}}

313: {4\pi D^2}e^{-\sigma(E)N_\rmscr{H}}

314: \int_0^1 2xdx\int_0^{2\pi}\frac{d\phi}{2\pi}\nonumber \\ &\times& \left\{ 

315: \frac{1}{k T_{p}}\sigma T^4_\rmscr{th}(\theta,\phi) \;

316: n[Ee^{-\Lambda_s};T_\rmscr{th}(\theta,\phi)]\;+\;

317: \frac{1}{k T_\rmscr{hs}}\sigma T^4_\rmscr{hs}(\theta,\phi) \;

318: n[Ee^{-\Lambda_s};T_\rmscr{hs}(\theta,\phi)]\right\}\;,

319: \label{eq:flux}

320: \end{eqnarray}

321: in units of phot cm$^{-2}$ s$^{-1}$ keV$^{-1}$. Note that the dependence 

322: on $\gamma$ comes in through Eqs. (1) -- (3).

323: 

324: Finally, the energy dependent pulsed fraction is defined by

325: \beq

326: Pf(E)=\frac{F^\rmscr{max}(E)-F^\rmscr{min}(E)}

327: {F^\rmscr{max}(E)+F^\rmscr{min}(E)}\;,

328: \label{eq:pf}

329: \eeq

330: where $F^\rmscr{max}(E)$ and $F^\rmscr{min}(E)$ are, respectively, the maximum

331: and minimum flux over a rotation period of the star.  The phase of the

332: pulsation, $Ph(E)$, is defined as the phase angle $\gamma$ at which the flux 

333: is maximum at energy $E$. 

334: 

335: \section{Application to PSR 1055-52, PSR 0656+14 and PSR 0630+18}

336: 

337: For each pulsar, we considered three types of models: a two-component

338: blackbody (of the type expected from iron atmospheres),

339: or a two-component light-element atmosphere, and for the

340: latter we considered two different types of opacities.  The emergent

341: spectrum from iron atmospheres is similar to a blackbody in its gross

342: properties (\cite{Raja97}) and we assume that it is moderately beamed

343: as discussed in \S 2.  The two-component light-element atmosphere is

344: substantially bluer than a blackbody and also exhibits a hard tail.

345: We have used two different semianalytic atmospheres of

346: \jcite{Heyl98atm} with opacities decreasing as $\nu^{-3}$ and

347: $\nu^{-1}$.  The former model is much bluer than a blackbody and

348: exhibits a strong hard tail.  

349: The latter is most appropriate to model magnetized light-element

350: atmospheres (\eg\ \cite{Pavl94}); it is intermediate between the

351: $\nu^{-3}$-model and the blackbody.

352: 

353: The effective temperatures (parameterized through $T_p$ and $T_{\rm

354: hs}$) and the size of the hot region (called $\beta$ in our paper)

355: correponding to the same effective blackbody temperatures are

356: different in the three models, and, to precisely determine them in

357: each case, we generated a simulated spectrum (using the software

358: XSPEC, \cite{A96}, and the $ROSAT$ PSPC detector) of each object,

359: using the temperature values of the double-blackbody fits and the

360: ratio between the areas of the two components given in the literature.

361: This spectrum was then fitted with the phase averaged spectrum of our

362: model\footnote{This is given by $\frac{1}{2\pi} \int_0^{2\pi}d\gamma

363: F(E;\gamma)$.}  to determine $T_p$, $T_{\rm hs}$ and $\beta$. These

364: values were then used to compute pulsed fractions and phases with some

365: choices of $n_1$ and $n_2$.

366: 

367: We found that an iron-type atmosphere, which has a rather mild degree

368: of beaming, was unable to reproduce the degree of modulation observed

369: in the softest part of the spectrum (dominated by the thermal

370: component\footnote{Note, however, that this result is dependent on the

371: assumption of a dipolar field. The value of the pulsed fractions can be

372: increased if a quadrupole component is added (\cite{Page96}).}). 

373: A light element atmosphere with an opacity $\propto\nu^{-3}$ was able

374: to reproduce reasonably well the pulsed fractions in the soft component

375: of the spectrum. However, the tail of the thermal component

376: is generally so strong that a

377: second hotter component was not statistically required by the data

378: in the case of PSR 0656 and PSR 1055, and it was much smaller

379: than the one required by the double blackbody fit in the case of Geminga\footnote{

380: Note that in the case of Geminga, similar results had been found by \jcite{MPM94}.}.

381: Such a model was unable to produce the sudden transition in the

382: behaviour of the pulsed fractions accompanied by a phase shift around

383: 0.5 keV which is observed in PSR 0656 and PSR 1055. 

384: Moreover, such models implied surface areas much larger than what would

385: be consistent with any reasonable equation of state for the neutron star. 

386: A two-component atmosphere model which is less blue ($\propto\nu^{-1}$) 

387: than the former required a second hotter component in the spectral fitting,

388: and was able to reproduce the behaviour of the observed pulsed fractions and

389: phase shifts. More details on each of the object that we considered are

390: given in the following. 

391: 

392: \subsection{PSR 1055-52}

393: 

394: This pulsar has a characteristic age $\tau=P/2\dot{P}\sim 5\times

395: 10^5$ yr, and a rotational energy loss rate of $\dot{E}\sim 3\times

396: 10^{34}$ erg/sec. {\em ROSAT} observations (\cite{Ogel93a}) clearly

397: show that the pulses have an energy dependent phase and pulsed

398: fraction, with a transition at $\sim 0.5$ keV. The pulsed fraction

399: below the transition is roughly constant at a value on the order of

400: 7-8\%, while after the transition it rapidly increases to a maximum of

401: 0.85 around 1 keV. The hard photons lead the soft ones by an angle

402: $\sim 120^\circ$.  A blackbody fit to the spectrum requires two

403: components, with temperatures $T_\rmscr{soft} \sim 8 \times 10^5$ K

404: and $T_\rmscr{hard}\sim3.7\times 10^5$ K, and a ratio between the 

405: areas of $A_\rmscr{soft}/A_\rmscr{hard}\sim 3\times 10^{-5}$

406: (\cite{Grei96}).  The angles $\xi$ and $\chi$ have been inferred by

407: \jcite{Malo90} to be both about $30^\circ$. With these values, and

408: $T_\rmscr{p}$, $T_\rmscr{hs}$ and $\beta$ calibrated on the phase averaged spectrum,

409: we show the derived pulsed fractions and phase for this pulsar in Figure

410: 1, having assumed $n_1=1$ and $n_2=2$.

411: 

412: \subsection{PSR 0656+14}

413: 

414: PSR 0656+14 was discovered in an {\em Einstein} satellite survey of

415: ultrasoft sources (\cite{1989ApJ...345..451C}). It has a relatively

416: young spin down age of $1.1\times 10^5$ yr and a rotational energy

417: loss rate $\dot{E}\sim 3.8 \times 10^{34}$ erg/sec. The combined {\em

418: ASCA} and {\em ROSAT} spectrum reveals two blackbody components with

419: $T_\rmscr{soft}\sim 8\times 10^5$ K, and $T_\rmscr{hard}\sim 1.5\times

420: 10^6$ K, and shows evidence that a power-law component is needed to

421: account for higher energy photons. The ratio of the hot polar cap

422: area to the neutron star surface area is $\sim 7\times 10^{-3}$

423: (\cite{Grei96}).

424: 

425: As for PSR 1055-52, both the phase and the pulsed fraction are energy

426: dependent. Below $\sim 0.5$ keV, the pulsed fraction stays a little under

427: $\sim 10\%$, and it increases modestly in the range $0.5-1$~keV

428: to around

429: 20-30 \% at 1 keV. The soft and the

430: hard components are shifted in phase by about $85^\circ$ with respect

431: to each other.  The angles $\xi$ and $\chi$ have been estimated by

432: \jcite{Malo90} to be both about $35^\circ$. With these values, and

433: $T_\rmscr{p}$, $T_\rmscr{hs}$ and $\beta$ derived from spectral fitting,

434: we show the pulsed fraction and phase that

435: our model predicts with the choice $n_1=0.8$ and $n_2=0.2$ in Figure 2.

436: 

437: \subsection{PSR 0630+18 (Geminga)}

438: 

439: Geminga was first observed as a high-energy $\gamma$-ray source with

440: the SAS-2 satellite in the 100 MeV band 

441: (\cite{Fich75}). It was

442: only in 1992 that it was observed in the X-ray band with {\em

443: ROSAT} (\cite{Halp92}).  Its period and spin-down rate (\cite{Bert92}; 

444: \cite{Herm92})

445:  yield a dynamic age of $3.4\times

446: 10^5$ yr, and a rotational energy loss rate $\dot{E} \sim 3.5\times

447: 10^{35}$ erg/sec. A spectral analysis of the {\em ROSAT} data made

448: by \jcite{Halp93} shows that the X-ray spectrum consists

449: of two blackbody components with $T_\rmscr{soft}\sim 5\times 10^5$ K,

450: and $T_\rmscr{hard}\sim 3\times 10^6$ K. Both components are modulated

451: at the pulsar rotation period, but the harder X-ray pulse leads 

452: the soft pulse by about $105^\circ$ in phase. Geminga is believed to

453: be an orthogonal rotator, with $\xi=\chi\sim 90^\circ$ (\cite{Malo90}). 

454: 

455: Unlike PSR 1055-52 and PSR 0656+14, the pulsed fraction observed in

456: Geminga shows a decrease with energy at low energies: the amplitude of

457: the pulsations in the PSPC channels 8-28 (i.e. roughly at energies

458: below 0.3 keV) is much larger than in channels 28-53 (roughly

459: corresponding to the energy range 0.3-0.5 keV): 33\% versus 20\%.

460: Figure 3 shows that a significant decrease of the pulsed

461: fractions in the 0.3-0.5 keV energy range can be reproduced with the

462: model described in \S 2: here we have taken $n_1=1.5$ and $n_2=3$,

463: while all other model parameters have again been calibrated from a fit

464: to the phase-averaged spectrum\footnote{Note that a similar behaviour

465: for the $Pfs$ in the softest part of the spectrum 

466: could be obtained with a more modest beaming (i.e. smaller $n_1$), but

467: a quadrupole component of the $B$ field added to the dipole (\cite{}Page96).}.  

468: The decrease in the pulsed fraction

469: occurs in the region where the harder component starts to overtake the

470: softer one. The hard component brings more photons at phase angles in

471: which the soft component has less photons, therefore leading to a

472: decrease of the intensity fluctuations produced by the thermal

473: component alone.  This effect happens in the same way also in the

474: other two examples that we showed, but due to the less intense beaming

475: of the harder component assumed there, the decrease is not so

476: pronounced.  Interestingly, a slight decrease in the pulsed fraction

477: before its rapid increase can be observed also in the data for these

478: other two pulsars (e.g. \cite{Ogel95}).

479: 

480: Our model for the phase shifts and pulsed fraction in Geminga is

481: similar to that proposed by \jcite{Halp93}.  \jcite{Page95b} later

482: argued that this picture could not work because the flux in the hard

483: component is more than an order of magnitude less than that in the

484: soft component in the $0.3-0.5$ keV energy range, where the decrease  

485: in the $Pf$ is observed. However, the

486: pencil-beaming reduces the flux of the soft component at the same

487: phase of the rotation where the fan-beaming of the hard component

488: increases its flux; therefore, although the phase averaged fluxes in

489: the two components are not comparable, the instantaneous fluxes 

490: at $\gamma \sim 90^\circ$ can be, 

491: and the hard component can begin to influence the soft one well before

492: it becomes dominant in the phase-averaged spectrum.  

493: 

494: \section{Summary and Discussion}

495: 

496: We have proposed a simple model that is able to account for properties

497: of the X-ray emission observed in the three middle aged pulsars PSR

498: 1055-52, PSR 0656+14 and Geminga within a unified framework.  The

499: emission in the $\sim 0.1-1$ keV energy range is believed to be due to

500: the combination of thermal radiation from the entire star and emission

501: from heated polar caps. We have argued that, while pencil beaming is

502: expected for the thermal, soft component, the hard component from the

503: polar caps is expected to be beamed into a fan, as discussed in the

504: literature for similar problems in other contexts.  We have shown that

505: such a model is able to account not only for the magnitude of the

506: pulsations in PSR 1055-52 and PSR 0656+14, but also for the phase

507: shifts between the soft and the hard component observed in all the

508: three objects.  We have shown that this same model is able to

509: reproduce also the decrease of the pulsed fraction observed in the

510: soft X-ray emission of Geminga, without requiring any special model

511: for the composition of the surface of this object. However, we have

512: found that the type and composition of the atmosphere plays an

513: important role. An atmosphere made of heavy elements is not able to

514: account for significant $Pfs$ (if only a dipolar field is assumed for

515: the thermal component), while a light element atmosphere with opacity

516: $\propto\nu^{-3}$ has a very pronounced tail which is able to account

517: for the all spectrum (up to about 1.5 - 2 keV) for PSR 0656 and PSR

518: 1055, without requiring a second, hotter component.  Such a model is

519: not able to account for the sudden increase of the $Pfs$ (accompanied

520: by a phase shift) that is observed for these two pulsars around 0.5

521: keV. In the case of Geminga, the implied size of the hot spots is too

522: small to produce any significant effect.  On the other hand, a light

523: element atmosphere with opacity $\propto\nu^{-1}$ requires a second,

524: hotter component to account for the overall spectrum, and, as we have

525: shown, such a model is able to account for the observed pulsed

526: fractions and phase shifts, if one allows the emission from the hot

527: spots to have a variable degree of beaming for the various objects.

528: We have argued that this this is plausible, as it depends on the

529: spatial and energy distribution of the bombarding particles, which is

530: likely to vary among the various objects.

531: 

532: In conclusion, we need to stress that the model we have presented here was

533: aimed at explaining features observed in the X-ray emission from

534: middle aged pulsars from a qualitative point of view.  A more

535: quantitative analysis, which includes detector and absorption effects

536: \footnote{For a perfect (diagonal) detector, absorption

537: does not affect the {\em energy-dependent} $Pfs$ (but it does

538: affect the $Pf$ over a finite energy bandwidth; \cite{PHH2000}); however, for a real

539: detector, absorption can affect also the energy-dependent $Pfs$ to

540: a certain extent (\cite{Page95}).}  

541: would not change any of the qualitative features reproduced here. On

542: the other hand, a thorough analysis, inclusive of

543: detailed predictions for the shape of the pulse profiles, would require

544: very detailed models for the atmosphere and for the temperature

545: distribution on the star, as well as much better data with which to 

546: compare the model.  We anticipate that forthcoming

547: observational data

548: from the {\em Chandra} and {\em XMM} missions

549: will soon make this endeavor feasible.

550: 

551: \acknowledgements{We thank Jonathan McDowell for support with use of

552:  the XSPEC software}.

553: 

554: %\bibliographystyle{jer}

555: %\bibliography{ns,physics,mine,gr}

556: \begin{thebibliography}{}

557: 

558: \bibitem[\protect{Arnaud~\protect\oyear 1996\protect\cyear}]{A96}

559: \newblock

560: Arnaud, K. A. 1996, in ASP Conf. Series 101, {\em Astronomical

561: Data Analysis Software and Systems V}, ed. G. Jacoby \& J. Barnes

562: (San Francisco: ASP), 17

563: 

564: \bibitem[\protect{Bertsch et~al.~\protect\oyear 1992\protect\cyear}]{Bert92}

565: Bertsch, D.~L. {\it et~al.} 1992,

566: \newblock {\em Nature,} {\bf 357}, 306.

567: 

568: \bibitem[\protect{Becker \& Trumper~\protect\oyear 1997\protect\cyear}]{BT97}

569: Becker, W. \& Trumper, J. 1997,

570: \newblock {\em A\&A,} {\bf 326}, 682

571: 

572: \bibitem[\protect{Chiu \& Salpeter~\protect\oyear 1964\protect\cyear}]{Chiu64}

573: Chiu, H.-Y. \& Salpeter, E.~E. 1964,

574: \newblock {\em Phys. Rev. Lett.,} {\bf 12}, 413.

575: 

576: \bibitem[\protect{{Cordova} et~al.~\protect\oyear

577:   1989\protect\cyear}]{1989ApJ...345..451C}

578: {Cordova}, F.~A., {Middleditch}, J., {Hjellming}, R.~M. \& {Mason}, K.~O. 1989,

579: \newblock {\em \apj,} {\bf 345}, 451.

580: 

581: \bibitem[\protect{Fichtel et~al.~\protect\oyear 1975\protect\cyear}]{Fich75}

582: Fichtel, C.~E. {\it et~al.} 1975,

583: \newblock {\em ApJ,} {\bf 198}, 163.

584: 

585: \bibitem[\protect{Greenstein \& Hartke~\protect\oyear

586:   1983\protect\cyear}]{Gree83}

587: Greenstein, G. \& Hartke, G.~J. 1983,

588: \newblock {\em ApJ,} {\bf 271}, 283.

589: 

590: \bibitem[\protect{Greiveldinger et~al.~\protect\oyear

591:   1996\protect\cyear}]{Grei96}

592: Greiveldinger, C. {\it et~al.} 1996,

593: \newblock {\em ApJL,} {\bf 465}, 35.

594: 

595: \bibitem[\protect{Halpern \& Holt~\protect\oyear 1992\protect\cyear}]{Halp92}

596: Halpern, J.~P. \& Holt, S.~S. 1992,

597: \newblock {\em Nature,} {\bf 357}, 222.

598: 

599: \bibitem[\protect{Halpern \& Ruderman~\protect\oyear

600:   1993\protect\cyear}]{Halp93}

601: Halpern, J.~P. \& Ruderman, M. 1993,

602: \newblock {\em ApJ,} {\bf 415}, 286.

603: 

604: \bibitem[\protect{Hermsen et~al.~\protect\oyear 1992\protect\cyear}]{Herm92}

605: Hermsen, W. {\it et~al.} 1992,

606: \newblock IAU Circular No. 5541

607: 

608: \bibitem[\protect{Heyl \& Hernquist~\protect\oyear

609:   1998\protect\cyeara}]{Heyl97analns}

610: Heyl, J.~S. \& Hernquist, L. 1998a,

611: \newblock {\em MNRAS,} {\bf 300}, 599.

612: 

613: \bibitem[\protect{Heyl \& Hernquist~\protect\oyear

614:   1998\protect\cyearb}]{Heyl98rxj}

615: Heyl, J.~S. \& Hernquist, L. 1998b,

616: \newblock {\em MNRAS,} {\bf 297}, L69.

617: 

618: \bibitem[\protect{Heyl \& Hernquist~\protect\oyear

619:   1998\protect\cyearc}]{Heyl98atm}

620: Heyl, J.~S. \& Hernquist, L. 1998c,

621: \newblock {\em MNRAS,} {\bf 298}, L17.

622: 

623: \bibitem[\protect{Heyl \& Hernquist~\protect\oyear

624:   2000\protect\cyear}]{Heyl2000}

625: Heyl, J.~S. \& Hernquist, L. 2000,

626: \newblock {\em MNRAS}, in press 

627: 

628: \bibitem[\protect{Malov~\protect\oyear 1990\protect\cyear}]{Malo90}

629: Malov, I.~F. 1990,

630: \newblock {\em Astron. Zh.,} {\bf 67}, 377.,

631: \newblock Sov. Astron, 34, 189

632: 

633: \bibitem[\protect{Meyer et al.~\protect\oyear 1990\protect\cyear}]{MPM94}

634: Meyer, R. D., Pavlov, G. G., \& Meszaros, P. 1994,

635: \newblock {\em \apj} {\bf 433}, 265

636: 

637: \bibitem[\protect{{Milgrom} \& {Salpeter}~\protect\oyear

638:   1975\protect\cyear}]{1975ApJ...196..583M}

639: {Milgrom}, M. \& {Salpeter}, E.~E. 1975,

640: \newblock {\em \apj,} {\bf 196}, 583.

641: 

642: \bibitem[\protect{{\"{O}}gelman~\protect\oyear 1995\protect\cyear}]{Ogel95}

643: {\"{O}}gelman, H. 1995,

644: \newblock in M.~A. Alpar, U. Kiziloglu \& J.~V. Paradijs (eds.), {\em The Lives

645:   of Neutron Stars}, p. 101, Kluwer, Dordrecht

646: 

647: \bibitem[\protect{{\"{O}}gelman \& Finley~\protect\oyear

648:   1993\protect\cyear}]{Ogel93a}

649: {\"{O}}gelman, H. \& Finley, J.~P. 1993,

650: \newblock {\em ApJL,} {\bf 413}, 31.

651: 

652: \bibitem[\protect{Page~\protect\oyear 1995\protect\cyear}]{Page95}

653: Page, D. 1995,

654: \newblock {\em ApJ,} {\bf 442}, 273.

655: 

656: \bibitem[\protect{Page \& Sarmiento~\protect\oyear 1996\protect\cyear}]{Page96}

657: Page, D. \& Sarmiento, A. 1996,

658: \newblock {\em ApJ,} {\bf 473}, 1067.

659: 

660: \bibitem[\protect{Page, Shibanov \& Zavlin~\protect\oyear

661:   1995\protect\cyear}]{Page95b}

662: Page, D., Shibanov, Y.~A. \& Zavlin, V.~E. 1995,

663: \newblock {\em ApJL,} {\bf 451}, 21.

664: 

665: \bibitem[\protect{Pavlov et~al.~\protect\oyear 1994\protect\cyear}]{Pavl94}

666: Pavlov, G.~G., Shibanov, Y.~A., Ventura, J. \& Zavlin, V.~E. 1994,

667: \newblock {\em A\&A,} {\bf 289}, 837.

668: 

669: \bibitem[\protect{Perna et~al.~\protect\oyear 2000\protect\cyear}]{PHH2000}

670: Perna, R., Heyl, J., \& Hernquist, L. 2000, 

671: \newblock {\em ApJL,} {\bf 538}, 159.

672: 

673: \bibitem[\protect{Rajagopal, Romani \& Miller~\protect\oyear 1997\protect\cyear}]{Raja97}

674: Rajagopal, M., Romani, R.~W., \& Miller, M.~C. 1997, 

675: \newblock {\em ApJ,} {\bf 479}, 347.

676: 

677: \bibitem[\protect{Shibanov et~al.~\protect\oyear 1995\protect\cyear}]{Shib95}

678: Shibanov, Y.~A., Pavlov, G.~G., Zavlin, V.~E. \& Tsuruta, S. 1995,

679: \newblock in H. B{\"{o}}hringer, G.~E. Morfill \& J.~E. Tr{\"{u}}mper (eds.),

680:   {\em Seventeenth Texas Symposium on Relativistic Astrophysics and Cosmology},

681:   Vol. 759 of {\em Annals of the New York Academy of Sciences}, p. 291, The New

682:   York Academy of Sciences, New York

683: 

684: \bibitem[\protect{Tsuruta~\protect\oyear 1964\protect\cyear}]{Tsur64}

685: Tsuruta, S. 1964,

686: \newblock {\em Ph.D. thesis}, Columbia University

687: 

688: \bibitem[\protect{{Zane}, {Turolla} \& {Treves}~\protect\oyear

689:   2000\protect\cyear}]{2000ApJ...537..387Z}

690: {Zane}, S., {Turolla}, R. \& {Treves}, A. 2000,

691: \newblock {\em \apj,} {\bf 537}, 387.

692: 

693: \bibitem[\protect{Zavlin, Pavlov \& Shibanov~\protect\oyear

694:   1996\protect\cyear}]{Zavl96}

695: Zavlin, V.~E., Pavlov, G.~G. \& Shibanov, Y.~A. 1996,

696: \newblock {\em A\& A,} {\bf 315}, 141.

697: 

698: \end{thebibliography}

699: \newpage

700: 

701: \begin{figure}[t]

702: \centerline{\epsfysize=5.7in\epsffile{fig1.ps}}

703: \caption{Predicted energy-dependent phase and pulsed fraction for

704: the X-ray emission from PSR 1055-52.}

705: \label{fig:1}

706: \end{figure}

707: 

708: \begin{figure}[t]

709: \centerline{\epsfysize=5.7in\epsffile{fig2.ps}}

710: \caption{Predicted energy-dependent phase and pulsed fraction for

711: the X-ray emission from PSR 0656+14.}

712: \label{fig:2}

713: \end{figure}

714: 

715: \begin{figure}[t]

716: \centerline{\epsfysize=5.7in\epsffile{fig3.ps}}

717: \caption{Predicted energy-dependent phase and pulsed fraction for

718: the X-ray emission from Geminga.}

719: \label{fig:3}

720: \end{figure}

721: 

722: 

723: 

724: \end{document}

725: 

726: 

727: 

728: 

729: 

730: 

731: 

732: 

733: