1: \documentclass{kluwer}
2: \usepackage{epsfig}
3: %\usepackage{graphicx}
4: %\boldmath
5: %\mag 1200
6: %\oddsidemargin -0.5pc % appropriate for A4, you may need to vary value %
7: %\evensidemargin -0.5pc % appropriate for A4, you may need to vary value %
8: %\topmargin -5pc % appropriate for A4, you may need to vary value %
9: %\topmargin -3pc
10: %\textwidth=14cm
11: %\documentclass{kluwer} % Specifies the document style.
12: %\usepackage{graphicx}
13: \newdisplay{guess}{Conjecture}
14:
15:
16: \begin{document}
17:
18: %---------------------------begin of definition
19:
20: \def\b12{PSR B1259-63}
21: \def\beq{\begin{equation}}
22: \def\eeq{\end{equation}}
23: %------------------------------------------
24: \def\a{\alpha}
25: \def\g{\gamma}
26: \def\s{\sigma}
27: \def\e{\varepsilon}
28: \def\th{\theta}
29: \def\om{\omega}
30: %-------------------------------------------
31: \renewcommand{\deg}{$^{\circ}$}
32: \newcommand{\degm}{^{\circ}}
33: \newcommand{\etal}{{et~al.$\,$}}
34: %-----------------------------end of definition
35:
36: \begin{opening}
37: \title{Non-pulsed emission\\ from the binary system \b12 }
38: \author{Maria \surname{Chernyakova}}
39: \author{Andrei \surname{Illarionov}}
40: \runningauthor{Maria A. Chernyakova, Andrei F. Illarionov}
41: \runningtitle{Non-pulsed emission from the binary system \b12}
42: \institute{Astro Space Center of P.N.Lebedev Physical Institute}
43: %\date{April 15, 1993}
44:
45: \begin{abstract}
46: \b12 is the only known binary system with a radio pulsar from which the
47: non-pulsed radio and X-ray emission was detected. The companion star in this
48: system is a Be star SS 2883. A rapidly rotating radio pulsar is expected to
49: produce a wind of relativistic
50: particles. Be stars are known to produce highly asymmetric mass loss. Due to
51: the interaction of the pulsar wind and the Be star wind the system of
52: two shocks between the pulsar and the Be star forms. In this paper we show
53: that the observed non-pulsed radio emission from the system is a result of
54: the synchrotron emission of the relativistic particles in the outflow
55: beyond the shock wave and that the non-pulsed X-ray emission is due to the
56: inverse Compton scattering of the Be star photons on this
57: particles. \end{abstract} \keywords{pulsars:individual:\b12}
58:
59:
60: \end{opening}
61:
62:
63:
64:
65: {\small
66: \section{Introduction}
67: \b12 is the only known binary system with a radio pulsar from which the
68: non-pulsed radio and X-ray emission was detected. The radio pulsar \b12
69: of spin period $P$=47.76ms was discovered in a high-frequency survey of
70: the southern Galactic plane. Spindown luminosity is
71: equal to $L_p=9\times10^{35}$erg/s. Pulsar timing confirmed by later
72: observations revealed that the pulsar is in a high-eccentricity ($e$=0.87),
73: long-orbital period ($T$=3.4year) binary system with a massive companion.
74: Its companion is a 10th mag Be star SS2883 of luminosity
75: $L_*$=5.8$\times$10$^4L_\odot$, estimated radius
76: $R_*\sim$(6-10)$R_\odot$, mass $M\sim$10$M_\odot$. The distance to the
77: Earth is about 2 kpc, the binary separation at
78: periastron is equal to 10$^{13}$cm.
79: Assuming a pulsar mass of 1.4$M_\odot$ the implied inclination angle of
80: the binary orbit to the plane of sky is 36$^\circ$. The longitude of periastron is equal to 138$^\circ$ (Johnston \etal 1996). Timing measurements have
81: shown that the disc of the Be star
82: is likely to be highly inclined to the orbital plane (Wex \etal, 1988).
83:
84:
85: The extensive radio observations of \b12 were carried out during the
86: 1994 and 1997 periastron passage at frequencies 1.5, 2.3, 4.8 and 8.4 GHz
87: (Johnston \etal 1996, 1999). It turned out that when the pulsar is far
88: from the periastron the pulsar emission is highly linear polarized and its
89: intensity is practically independent on the pulsar orbital position. But
90: as the pulsar approaches to the periastron the properties of the pulsar
91: emission start to change. The depolarization of pulsed emission, the
92: increase of the dispersion measure and the absolute value of the rotation
93: measure near the periastron occurs. The pulsar was not detected in 1.5 GHz
94: data on 1993 December 20 (20 days before the periastron) and reappeared only on
95: 1994 February 4 (24 days after the periastron). During the 44 days eclipse no
96: pulse emission was detected in extensive observations at 1.5 and 8.4 GHz.
97: In the works of Lipunov \etal (1994)
98: and Johnston \etal (1996) it was shown that the observed eclipse is due to
99: the free-free absorption of the pulsar emission in the Be star disk.
100:
101: During the
102: 1994 and 1997 periastron passage the observations of the unpulsed
103: radio emission coming from the system were made using the Molongo Observatory
104: Syn\-thesis Telescope (MOST) and the Australia Telescope Compact Array
105: (ATCA) at five frequencies between 0.84 and 8.4 GHz (Johnston \etal 1999).
106: The first significant data at all frequencies were received 21 day before
107: the periastron. Since that time the intensity of the non-pulsed emission by
108: and large gradually increases until the 20th day after the periastron. That
109: day it reaches its maximum (57 mJy at 1.4 GHz) and then it starts to
110: decline. The source remained significant in the ATCA data until about 60
111: days after the periastron and until about hundred days after the periastron
112: in the MOST data. At times when the pulsar emission is eclipsed, there is
113: no evidence of polarized emission in any of the observations in either 1994
114: or 1997.
115:
116: \begin{figure} %proc1.eps
117: \centering\epsfig{file=proc1.eps,width=12cm, height=20cm}
118: \vspace{-10cm}
119: \caption {The geometry under consideration. Point $P$ marks the position
120: of the pulsar, point $Be$ - the position of the Be star. Curves $S_1$,
121: $S_2$ represent the shock waves, curve $c$ - the contact surface. Curve
122: $U$ is a curve of the constant density (see text). Points $A$ and $B$ mark
123: the intersection of the curve $U$ with the shocks $S_1$ and $S_2$,
124: point $C$ is the point of the shock $S_2$ intersection with the binary
125: axis. $ED$ is a small arc centred at the point $P$.}
126: \end{figure}
127:
128:
129:
130: Since its discovery the PSR B1259-63 system was observed several
131: times by X-ray instruments
132: (see review in Tavani \etal 1997, Hirayama \etal 1999).
133: It was shown that
134: the X-ray spectrum of the system is consistent with a power-law of photon
135: index $\sim-1.7$. No significant X-ray pulsations with the pulsar spin
136: period were detected.
137:
138: The aim of this paper is to explain the origin of the non-pulsed radio and
139: X-ray emission. It is well-known that pulsars lose rotational energy in
140: the form of relativistic MHD winds (Michel 1969, Arons 1992). The wind
141: carries energy flux in the form of electromagnetic fields and the kinetic
142: energy of the relativistic electrons and positrons. The energy flux in the
143: pulsar wind ${\cal F}_w$ at distance $r$ well over the light cylinder is equal to
144: \begin{equation}
145: {\cal F}_w=\frac{L_p}{4\pi r^2}=\frac{cB_w^2}{4\pi\sigma}.
146: \label{b0}
147: \end{equation}
148: $B_w$ is a magnetic field in the wind, $\sigma\ll 1$
149: is a pulsar wind magnetization, i.e. the upstream ratio of electromagnetic
150: energy density and particle kinetic energy density. For the Crab Nebula
151: $\sigma\sim0.005$ (Kennel\& Coronity, 1984).
152:
153: Be stars are well-known to be the source of the strong highly
154: anisotropic matter outflow. Both a dilute polar wind and a denser
155: equatorial disk have been invoked to reconcile models for IR, UV and
156: optical observations of Be stars (Waters \etal 1988).
157: Due
158: to the interaction of the pulsar and Be-star winds the system of two shocks arises and
159: the relativistic particles flowing at first radially from the pulsar turn
160: after passing the shock and start to outflow along the contact surface.
161: In the paper of Tavani \etal (1997) the non-pulsed X-ray emission was explained as a result of the synchrotron emission of the highly relativistic pulsar wind particles with Lorentz factor $\gamma \sim 10^6$. In this model the origin of the unpulsed radio emission is unclear. Here we propose a model in which Lorentz factor is supposed to be much more moderate, $\gamma \sim 10 $, and show that in this case the unpulsed radio emission is a result of the
162: synchrotron radiation of the relativistic particles in the outflow beyond
163: the shock and that the non-pulsed X-ray emission is due to the
164: inverse Compton scattering of the Be star photons on this particles.
165:
166: \section{Colliding winds}
167:
168: In order to calculate the unpulsed emission coming from \b12
169: it is necessary to know
170: the structure of the shock waves arising in the system due to the interaction
171: of the relativistic pulsar wind and a non-relativistic Be star wind.
172:
173: The structure of the colliding winds in the system \b12 is very complicate.
174: But the origin of the unpulsed emission can be understood with
175: the help of the rather simple model.
176: For the simplicity we suppose that the polar wind of the
177: Be-star occupies all the region outside the disk and that in that region it
178: is homogeneous. The pulsar wind we suppose to be spherically symmetric. The
179: observations of the inner regions of the Crab Nebula (Hester \etal 1995)
180: approves this assumption. Then for the most part of the pulsar orbit the
181: geometry of the problem reduces to the interaction of the two spherically
182: symmetric winds.
183:
184:
185: The interaction of the relativistic and non-relativistic winds is still
186: poor investi\-gated but it seems that it is closely analogous to the case
187: of the interaction of two non-relativistic winds (e.g. Melatos \etal 1995).
188: For the non-relativistic case it was
189: shown by Lebedev \& Myasnikov (1990) that if the dynamical pressures of the two winds are of the same order then the form of the shocks converges to a hollow cone at
190: distances bigger then the distance between the pulsar and a contact
191: surface.
192:
193: Equating the dynamical pressures
194: of the winds we find under the assumption of the winds isotropy the
195: distance to the contact surface from the pulsar:
196: \beq
197: r_c=a\frac{\sqrt{\a}}{1+\sqrt{\a}}\;\;\;\;\; , \;\;\;\;\;\;
198: \a=\frac{L_p}{cv_0\dot{M}},
199: \label{rc}
200: \eeq
201: $a$ is a
202: binary separation, $\dot{M}$ is a Be star mass outflow rate, $v_0$ is a Be
203: star polar wind termination velocity, $c$ is a light velocity. In the paper
204: of Snow (1982) it was found out that the rate of the Be star mass loss
205: is connected with its luminosity. From that dependence it follows that
206: the rate of SS2883 mass loss should be about $\dot{M}=$10$^{-8}\dot{M}_{-8}M_{\odot
207: }$/year. The typical terminal velocity of the Be star polar wind is about
208: $v_0=10^8v_8$cm/c (Snow 1981). Thus we find that for \b12
209: $\alpha=0.45/v_8\dot{M}_{-8}$. Hence the winds have comparable values of
210: the dynamical pressure and the contact surface crosses the binary axis
211: somewhat closer to the pulsar then to the Be star.
212:
213:
214: In the beginning both the relativistic and non-relativistic winds are
215: radial but after passing the shock winds turn and start to flow along
216: the contact surface.
217: Both winds are highly supersonic and thus both shocks
218: arising in the system are strong
219: and MHD shock conditions states that the normal component of the particles
220: velocity after the shock is lowered by a factor 3 and 4 for relativistic
221: and non-relativistic particles correspondingly (Landau\&Lifshitz 1986).
222: The big difference in the velocities of the winds at
223: different sides of the contact surface can lead to the growth of
224: instability and the two winds will be macroscopically mixed between the
225: shocks. In this case the region between the shocks will be filled with the
226: two-phase outflow composed of volumes with relativistic and volumes with
227: non-relativistic particles. Then the heavy non-relativistic wind slows
228: down the volumes filled by the relativistic electrons and positrons and
229: they acquire essentially non-relativistic hydrodynamic drift velocity $v_d$
230: along the shock while the energy of electrons and positrons does not
231: changes significantly. The non-relativistic plasma in its turn accelerates and acquire the velocity of the order of $v_d$. The analysis of the
232: observational data leads to the conclusion that the similar
233: situation occurs in systems G70.7+1.2
234: (Kulkarni \etal 1992) and LSI 61\deg303 (Maraschi\&Treves 1981) and
235: that the observed emission from these systems is generated
236: in the regions filled with the two-phase flow.
237:
238: \section{Generation of the non-pulsed X-ray emission}
239: The observed power-law spectrum of the non-pulsed X-ray emission coming
240: from the system \b12 suggests a power-law energy distribution of the
241: relati\-vistic particles in the outflow after the shock
242: $\frac{dN_{e^\pm}}{ d\varepsilon }=K_e\varepsilon ^{-s}$,
243: $\varepsilon>\varepsilon_{min}$
244: (Chernyakova \& Illarionov 1999).
245: Such a distribution can be either the result of the intrinsic power law
246: distribution of the electrons and positrons in the pulsar wind (e.g.
247: Tademaru 1973; Lominadze \etal 1983), or in the case of the particle
248: acceleration at the shock (e.g. Gallant \etal 1994).
249:
250: The soft photons, emitting by the Be-star would scatter on the relativistic
251: particles, outflowing with the drift velocity $v_d$.
252: The scattered photons form a wide spectrum from the X-ray band up to the
253: gamma-ray band. Applying the results of the work (Chernyakova \& Illarionov
254: 1999)\footnote{In that work there is a misprint - in (27) the value of
255: $v_d=10^9$cm/s was taken.} to the system PSR B1259-63 we estimate that for $\varepsilon
256: _{min}=10mc^2, s=2.4, v_d=10^9$cm/s
257: during the periastron passage
258: the photon flux density of the unpulsed
259: X-ray emission from the system at Earth should be about
260: \begin{equation}
261: F\sim 5\times 10^{-3}(\frac \varepsilon {100keV})^{-1.7}ph/s/cm^2/MeV,
262: \end{equation}
263: while the observations give at 100 keV the value\\ $F_{obs}=(2.8\pm 0.7)\times
264: 10^{-3}$ph/s/cm$^2$/MeV (Grove \etal 1995).
265:
266: \section{Non-pulsed radio emission as a result of the synchrotron emission
267: of the electrons of a pulsar wind}
268: Under the assumption of the relativistic particles power law energy distribution
269: $dn_{e^\pm}=K_e\varepsilon^{-2.4}d\varepsilon$,
270: $\varepsilon>10mc^2$, (the
271: parameter $K_e$ characterize the density of relativistic particles),
272: the flux
273: density detected at the Earth from the synchrotron emission on a frequency
274: $\nu$ ( throughout the text $\nu$ is measured in GHz) from the small volume
275: $dV=dldS=dV_{39}*10^{39}$cm$^3$ (with the side $dl$ along the line of
276: sight) with a magnetic field $B$ is equal to (Corchak, Terleckiy 1952)
277: \begin{equation}
278: dI_E=\frac{J_\nu dV}{4\pi D^2}=3.23\times
279: 10^5\frac{K_e}{4\pi}\frac{dV_{39}}{D_2^2}(B\sin{\chi})^{1.7}\nu^{-0.7}mJy,
280: \label{inu}
281: \end{equation}
282: where $\chi$
283: is an angle between the magnetic field and the direction to
284: the observer, $D=D_2*2$kpc is a distance to the system and $J_\nu$ characterizes the energy emitted from a unit volume per a unit time.
285:
286: The coefficient of self-absorption is equal to (Ginzburg \etal 1969)
287: \begin{equation}
288: \mu=1.1\times 10^{-8}K_eB^{2.2}\nu^{-3.2}cm^{-1}
289: \label{pogl}
290: \end{equation}
291:
292: In order to calculate the flux density with formulas (\ref{inu}),
293: (\ref{pogl}) it is necessary to know the density of the relativistic
294: particles in the outflow $K_e$ and the magnetic field $B$.
295: To calculate the parameters of the relativistic outflow at an arbitrary
296: point $I$ located between the two shocks we use a simplifying assumption
297: of the parameters constancy across the
298: shock (along the surface $U$, Fig.1).
299:
300: \subsection{The relativistic particles density}
301: The density of the relativistic particles at the surface
302: $U$ can be estimated from energy conservation law. Let us consider
303: contour $ABCDE$ (Fig.1). The energy flow through the closed surface arising
304: from the rotation of the contour $ABCDE$ about the binary axis is equal to
305: zero. Equating the inflowing flux density $L_p\Omega/4\pi$ to
306: outflowing one $K_eS_Uv_d/0.4\epsilon_{min}^{0.4}$ we find the parameter $K_e$:
307: \begin{equation} K_e=\frac{0.4
308: L_p\Omega\epsilon_{min}^{0.4}}{4\pi S_Uv_d},
309: \label{ke}
310: \end{equation}
311: here $S_U$ is
312: the surface $U$ area and $\Omega=2\pi(1-\cos\zeta)$, where
313: $\zeta$ is an angle $\widehat{APC}$ (Fig 1). Far from the pulsar
314: $S_U=2\pi R_U^2(\cos\theta_1-\cos\theta_2)$,
315: $R_U$- is a distance from the Be star to the surface $U$.
316:
317: \subsection{Magnetic field}
318: At distances much bigger then the radius of the light cylinder the
319: magnetic field in the pulsar wind may be considered as a toroidal one.
320: From (\ref{b0}) it follows that at a distance $r$ from the pulsar
321: it is equal to $B_w=\frac{1}{r}\sqrt{\frac{\sigma L_p}{c}}$. To calculate
322: the structure of a magnetic field in the outflow we use the equation of the
323: frozen magnetic field that in the stationary case have form (Lundquist
324: 1951) \begin{equation} rot[\vec{v} \times \vec{B}]=0. \label{frozen}
325: \end{equation}
326: Let us introduce the cylindrical coordinate system ($\eta,\varphi,z$) with
327: $z$ along the binary axis, chosen so that $\varphi_{obs}=0$.
328: Assuming that the outflow is
329: azimuthal symmetrical about the binary axis, that is $v_\varphi=0$,
330: we find from (\ref{frozen}):
331: \begin{equation}
332: B_{\varphi}=\sqrt{\sigma cL_p}
333: \frac{\zeta\sqrt{\cos^2\a+\sin^2\a\cos^2(\varphi-\varphi_p)}}
334: {v_d\int^{\theta_2}_{\theta_1}\frac{\eta_U}{\sin{\theta}}d\theta},
335: \label{b2} \end{equation} the integration in the denominator is done along
336: the surface $U$ at a constant $\varphi$, $\eta_U$ is a distance between
337: the elements of the surface $U$ and the binary axis,
338: vector $\vec{e_p}$ with the coordinates
339: $(\sin\a,\varphi_p,\cos\a)$ in our cylindrical coordinate system is aligned
340: with the pulsar axis.
341:
342: In the degenerate case of the coincidence of the binary axis and the axis
343: of the pulsar ($\alpha=0$) the magnetic field has only $\varphi$-component.
344:
345:
346:
347: In general case the magnetic field in the outflow also have a component
348: $B_d$ along the drift velocity and a component $B_U$ along the surface $U$.
349: The condition of the frozen
350: magnetic field suppose that the variation of $B/n_e$ at a volume with a
351: given particles in a course of motion is proportional to the stretching of a
352: magnetic field line passing through this volume (Lundquist 1951). With the
353: help of (\ref {ke}) we find that $B_\varphi$ and $B_U$ are proportional to
354: $r^{-1}$ and that $B_d\sim r^{-2}$. However if the mixing of the two winds
355: takes place then the direction of the component perpendicular to
356: $\vec{e}_\varphi$ changes all the time in the course of motion, this
357: component damps faster then the $B_\varphi$ and at a large distance from
358: the pulsar the $\varphi$-component of the magnetic field dominates.
359:
360: If the mixing of the wind doesn't occur, then far from the pulsar,
361: in the
362: region where the form of the shock waves $S_1$, $S_2$ is approximately a
363: cone, $B_d$ is small and we find from (\ref{frozen}) the value of $B_U$:
364: \beq
365: B_U=
366: %\frac{\Delta\th\frac{\partial
367: %B_\varphi}{\partial \varphi}}{\sin{\th_2}-\sin{\th_1}}=
368: -\frac{\sqrt{\s cL_p}}{\sin{\th_2}-\sin{\th_1}}\frac{\zeta}{R_U v_d}
369: \frac{\sin^2\a\cos(\varphi-\varphi_p)\sin(\varphi-\varphi_p)}
370: {\sqrt{\cos^2\a+\sin^2\a\cos^2(\varphi-\varphi_p)}}
371: \label{bth}
372: \eeq
373:
374:
375:
376:
377: \subsubsection{The intensity of the non-pulsed radio emission}
378: %The bulk of intensity comes from the regions with the optical depth about
379: %unity. According to (\ref{pogl}), (\ref{ke}) the optical depth is unity
380: %at a frequency $\nu_{cr}$
381: %\begin{equation}
382: %\nu_{cr}=10.4\left[\frac{l_{13}}{S_{26}v_{d9}}
383: %\frac{\varepsilon_{min}}{10mc^2}
384: %\frac{\Omega}{2\pi}B^{2.2}\right]^{\frac{1}{3.2}}GHz.
385: %\label{nu}
386: %\end{equation}
387: %Here $l=l_{13}*10^{13}$cm is the width of the layer and
388: %$S_U=S_{26}*10^{26}$cm$^2$ is a surface $U$ area. Thus the self-absorption
389: %is important at frequencies of interest.
390: With the help of formulas (\ref{inu}), (\ref{pogl}), (\ref{ke}),
391: (\ref{b2}) and (\ref{bth}) we can calculate the resulting synchrotron emission
392: from the
393: system. The differential equation describing the change of the flux
394: density $I_\nu$ along the line of view crossing the emitting region is
395: \beq
396: \frac{dI_\nu}{dl}=J_\nu(l)-I_\nu(l)\mu(l),
397: \label{deq}
398: \eeq
399: where $l$ is a coordinate along the line of sight. The total emission that
400: would be observed at the Earth is equal to:
401: \beq
402: I_E=\frac{1}{4\pi D^2}
403: \int\int_{l_0}^{l_e}J_{\nu}(l)\exp\left[-\int_{l}^{l_e}\mu(l_1)dl_1\right]
404: dldS.
405: \label{ieq}
406: \eeq
407: To calculate the observed flux we should choose the limits of integration
408: $l_0$ and $l_e$ so that this intercept totally covers that part of the line
409: of sight where the emission is generated and to integrate over the $l$
410: and over the area in the plane perpendicular to the line of sight
411: chosen so that it embraces all the lines of sight crossing the emission
412: region.
413:
414: Figure 2 represents the spectra calculated for different parameters in the case
415: of a pulsar with $\alpha=0$ located at a position
416: 30 days before the periastron passage. In our
417: calculations we follow the work of Melatos \etal (1995) and approximate the
418: form of the shock waves $S_1$, $S_2$ with hyperboloids with half-opening angles
419: $\theta_1, \theta_2$. It turns out that the bulk of radiation comes from
420: the regions away from the pulsar, where the drift velocity depends only
421: slightly on the position along the layer and for the simplicity we treat it
422: as a constant. The separation of the region generating the bulk of the
423: observed emission from the pulsar also leads to the weak dependence of the
424: result on the form of the shock waves in the vicinity of the pulsar.
425: For the comparison the observational spectra that were received by Johnston \etal (1999) 3 days before the periastron passage and 16 days after are also shown.
426:
427: \begin{figure}[t] %proc2.eps
428: %\centering\epsfig{file=proc2a.eps,width=12cm, height=12cm}
429: \epsfig{file=proc2a.eps,width=12cm, height=12cm}
430: \vspace{-5cm}
431: \caption{ The model spectra of the unpulsed radio emission coming from
432: the system 30 days from the periastron. In calculations the pulsar axis
433: was assumed to be directed along the binary axis.
434: 1.$v_d=c/15$, $\sigma=5\times 10^{-3}$, $\th_1=10\degm$, $\th_2=70\degm$
435: 2.$v_d=c/10$, $\sigma=5\times 10^{-3}$, $\th_1=10\degm$, $\th_2=70\degm$
436: 3.$v_d=c/10$, $\sigma=10^{-3}$, $\th_1=10\degm$, $\th_2=70\degm$
437: 4.$v_d=c/3$, $\sigma= 10^{-3}$, $\th_1=20\degm$, $\th_2=55\degm$.
438: Curves with error bars are the spectra observed from the system 3 days
439: before the periastron and 16 days after (Johnston \etal 1999).}
440: \end{figure}
441:
442:
443: Figure 2 shows that the flux density of the emission highly depends on the
444: value of the drift velocity $v_d$ and the magnetization parameter
445: $\sigma$.The lower the drift velocity is and the higher is the value of
446: $\sigma$ the more intensive is the resulted emission.
447:
448: In the case of the arbitrary position of the pulsar axis the estimation of the
449: value of the magnetic field along the layer is rather difficult. However as
450: we saw in the previous section if the mixing occurs then at large distances
451: from the pulsar the $\varphi$-component of the magnetic field dominates and
452: its value is close to the one in the degenerate case. The wide range of
453: the results received in the case of a pulsar with an axis along the binary
454: axis by variation of parameters shows that observed radio emission can
455: be explained with a synchrotron mechanism.
456:
457: \subsubsection{Polarization properties}
458: The degree of polarization $\Pi$ of the relativistic electrons with
459: the power-law energy distribution
460: $dn_{e^\pm}=K_e\varepsilon^{-2.4}d\varepsilon$
461: is given by (Corchak, Terleckiy 1952):
462: \beq
463: \Pi=\frac{\sqrt{{Q}_E^2+{U}_E^2}}{{I}_E}
464: \label{pol}
465: \eeq
466: Here $I_E$ is given by (\ref{ieq}) and the formulas for $Q_E$ and $U_E$ can
467: be received from (\ref{ieq}) by a substitution of
468: $0.72J_\nu(l)\cos 2\tilde{\chi}$ and $0.72J_\nu(l)\sin 2\tilde{\chi}$
469: for $J_\nu(l)$ correspondingly. $\tilde{\chi}$ is an angle between the arbitrary fixed
470: direction in the plane perpendicular to the line of sight and the direction
471: along the $\vec{e}_1=\vec{B}\times\vec{e}_o$, where $\vec{e}_o$ is a unit
472: vector along the line of sight.
473:
474: On Figure 3 the dependence of the degree of polarization on the angle
475: between the line of sight and the binary axis is shown. This dependence was
476: calculated according to (\ref{pol}) for the different positions of the
477: pulsar axis. It can be seen from the Figure 3 that the observed unpulsed radio emission should be highly polarized. However in formula (\ref{pol})
478: the effect of the Faraday rotation was not taken into account. The emission
479: is totally depolarized if the strong Faraday rotation is taken place in the
480: emission region (Burn, 1966): \beq
481: \Delta\varpi=\frac{0.236}{\nu^2}\int nB\cos\chi dl_{13}\gg 1.
482: \label{depol} \eeq Here $\nu$ is measured in GHz, $B$ in Gs , $n$ in
483: cm$^{-3}$ and $l_{13}$ in 10$^{13}$cm.
484:
485: If beyond the shock wave the winds stirring takes place, then the
486: electrons density in (\ref{depol}) is the density of
487: non-relavistic electrons $n_{nr}$ and the magnetic field is a magnetic
488: field of the Be-star $B_{Be}$. In the outflow $B_{Be}$ is about 1G and
489: $n_{nr}$ can be estimated from the particles conservation law
490: \beq
491: n_{nr}=\frac{\dot{M}(1-\cos\th_2)}{4\pi m_p v_{nr} R_U^2 (\cos\theta_1-cos\theta_2)}= 3\times 10^5\frac{\dot{M}_{-8}(1-\cos\th_2)}{v_{nr9}R_{U13}^2(\cos\theta_1-cos\theta_2)}cm^{-3}.
492: \label{nnr}
493: \eeq
494: Here $v_{nr}=v_{nr9}10^{9}$cm/s is a drift velocity of the non-relativistic
495: plasma beyond the shock wave and $\dot{M}$= $\dot{M}_{-8}M_\odot$/year is the rate of Be star mass
496: loss. $R_U=R_{U13}10^{13}$cm is a distance from the Be-star to the surface U.
497: The value of $n_{nr}$ varies along the shock wave but in the region of
498: interest it is, as it can be seen from (\ref{nnr}), high and depolarization takes place. If the mixing
499: doesn't occur and shock waves have such a form that the line of
500: sight doesn't pass through the region filled with the non-relativistic
501: electrons then from (\ref{ke}), (\ref{b2}), (\ref{bth}) and (\ref{depol})
502: it follows that $\Delta\varpi\sim1$, the effect of the Faraday rotation is
503: not strong and the radio emission would be polarized. Unfortunately
504: the degree of polarization in this case is not measured yet as the
505: intensity of the non-pulsed radiation in this case is low (see
506: Fig.2).
507:
508:
509:
510: \begin{figure}[t] %proc3.eps
511: \centerline{\epsfig{file=proc3a.eps,width=6cm, height=9cm}
512: \epsfig{file=proc3b.eps,width=6cm, height=9cm}}
513: \vspace{-2cm}
514: \caption{ The dependence of the degree of polarization on
515: the angle $\theta_{obs}$ between the line of sight and the binary axis. Angles
516: $\alpha$ and $\varphi_p$ define the position of the pulsar axis relative to the binary axis and the observer (see text). It was
517: calculated
518: under the assumption that the effect of the Faraday rotation is
519: negligible.}
520: \end{figure}
521:
522:
523:
524:
525: \section{The light curve}
526: The observations of the \b12 show that far from the periastron there is
527: no evidence for any
528: unpulsed radiation in the off-pulse bins down to a limit of several mJy.
529: The examination of the synchrotron emission of the pulsar wind relativistic
530: particles in the outflow beyond the shock leads us to the conclusion that
531: the drift velocity of the relativistic particles is high and there is no
532: mixing of the relativistic and non-relativistic winds beyond the shock.
533: According to our analysis the arising unpulsed radio emission is polarized
534: with the degree of the polarization about 30\%.
535:
536: The
537: detectable non-pulsed
538: radio emission from the system appears only 20 days before the
539: periastron passage and approximately at the same time the
540: pulsed emission disappears. It testifies that at this time the pulsar
541: crosses the disk of Be-star first time and the pulsed emission is absorbed
542: by the matter of a disk. The non-pulsed radio emission is generated rather
543: far from the pulsar and crosses the disk at its low-density part. Thus the
544: free-free absorption of the non-pulsed emission by a disk matter is
545: negligible.
546:
547: The interaction with the pulsar leads to the partial destruction of the
548: Be star disk (Ivanov \etal 1998). The matter ejecting from the disk
549: increases the instability in the outflow beyond the shocks leading to the
550: mixing of the relativistic and non-relativistic winds and to the
551: deceleration of the relativistic particles drift velocity. As it can be
552: seen from the Figure 2 the decreasing of the drift velocity leads to the
553: increase of the intensity of the generated non-pulsed radio emission. The
554: observed intensity can be explained with the drift velocity about $v_d\sim
555: c/10.$ The mixing of the winds leads to the increase of the electron
556: density on the line of sight and the emission becomes unpolarized.
557: The free-free optical depth along the line of sight is given by
558: $\tau_{ff}=5.7\times 10^{-15}\nu^{-2}T_{e5}^{-3/2}\int n_e^2dl_{13}$ ($\nu$
559: is measured in GHz, $l_{13}$ in 10$^{13}$cm, $n_e$ in cm$^{-3}$ and $T_{e5}$ in
560: $10^{5}$K). From (\ref{nnr}) we see that $\tau_{ff}\ll 1$
561:
562: The observed maximum of the unpulsed radio emission intensity
563: occurs at the time of the second pulsar intersection with the Be
564: star disk. After that the ejection
565: of the matter from the disk continues for some time but the father the
566: pulsar is from the disk the less is the influence of the ejected matter on
567: the winds outflow properties and at some moment the winds are not mixing
568: anymore. The unpulsed radio emission becomes again polarized
569: but weak.
570:
571: By this means we show that the observed properties of the non-pulsed
572: radio\-emission from the system \b12 can be explained by the synchrotron
573: emission of the pulsar wind relativistic particles in the outflow beyond
574: the shock wave and that the non-pulsed X-ray emission is due to the
575: inverse Compton scattering of the Be star photons on this
576: particles. }
577:
578: \acknowledgements
579: Authors are grateful to A.V.Myasnikov for the helpful remarks.
580: We acknowledge the valuable comments of the anonymous referee.
581: This work was supported by RFBR grant
582: 97-02-16975
583:
584: {\footnotesize \begin{thebibliography}{99}
585:
586:
587:
588: \bibitem{a1} Arons J., 1992, in Hawkins T., Rankin J., Gil J., eds Proc.
589: IAU Colloq. 128, ''The magnetosphere structure and emission mechanisms of
590: radio pulsars'', 56.
591:
592: \bibitem{p1} Burn B.J., 1966, MNRAS, 133, 67
593:
594: \bibitem{d16} Chernyakova M.A., Illarionov A.F., 1999, MNRAS, 304, 359
595:
596: \bibitem{d30} Corchak A.A., Terleckiy Ya.P., 1952, JETP (in russian), 22,
597: 507
598:
599:
600: \bibitem{p2} Gallant Y.A., Arons J., 1994, Ap.J., 435, 230
601:
602: \bibitem{d31} Ginzburg V.L., Syrovatskiy S.I.,
603: 1969, Ann.Rev.Astron.Astrophys. 7, 375
604:
605: \bibitem{a7} Grove J.E, Tavani M., Purcell W.R., Johnson, W.N., Kurfess
606: J.D., Strickman M.S., Arons J., 1995 ApJ, 447, L113.
607:
608:
609: \bibitem{d21} Johnston S., Manchester R.N., Lyne A.G., D'Amico N.,
610: Bailes M., Gaensler B.M., Nikastro L.,
611: 1996, MNRAS, 279, 1026
612:
613: \bibitem{d20} Johnston S., Manchester R.N., McConnell D.,
614: Campbell-Wilson D., 1999, MNRAS, 302, 277
615:
616: \bibitem{ba1} Hirayama M., Cominsky L.R., Kaspi V.M.,
617: Nagase F., Tavani M., Kawai N., Grove J.E.,
618: 1999 to be published in Ap.J., astro-ph/9905172
619:
620: \bibitem{p3} Hester J.J., \etal 1995, Ap.J., 448, 240
621:
622: \bibitem{p4} Ivanov P.B., Igumenshchev I.V., Novikov I.D., 1998, Ap.J., 507,
623: 131
624:
625: \bibitem{d5} Kennel C.F., Coronity F.V., 1984, Ap.J, 283, 694
626:
627: \bibitem{p5} Kulkarni S.R., Vogel S.N., Wang Z., Wood D.O.S., 1992,
628: Nature, 360, 139
629:
630:
631: \bibitem{b12a} Landau L.D. and Lifshitz E.M., "Hydrodynamic",
632: NewYork: Pergamon Press, 1975
633:
634: \bibitem{p6} Lebedev M.G., Myasnikov A.V., 1990, Fluid Dynamics, 25, 629
635:
636: \bibitem{p7} Lipunov V.M., Nazin S.N., Osminkin E.Yu., Prokhorov M.E., 1994,
637: A\&A, 282, 61
638:
639: \bibitem{p8} Lominadze J.G., Machabeli G.Z., Usov V.V., 1983, Ap\&SS,
640: 90, 19
641:
642: \bibitem{p9} Lundquist S., 1951, Phys.Rev., 83, 307
643:
644: \bibitem{b01a} Maraschi L., Treves A., 1981, MNRAS, 194, 1
645:
646: \bibitem{d22} Melatos A., Johnston S., Melrose D.B., 1995, MNRAS, 275, 381
647:
648: \bibitem{d1} Michel F.C., 1969, Ap.J., 158, 727
649:
650: \bibitem{p11} Snow T.P., 1981, Ap.J., 251, 139
651:
652:
653: \bibitem{p12} Snow T.P., 1982, Ap.J., 253, L39
654:
655: %\bibitem{d24} Stevens I.R., Blondin J.M., Pollock A.M., 1992, Ap.J.,
656: %386, 265
657:
658: \bibitem{p13} Tademaru E., 1973, Ap.J., 183, 625
659:
660: \bibitem{b23} Tavani M.,Arons J., 1997, ApJ 477,439
661:
662:
663: \bibitem{p14} Waters L.B.F.M., Taylor A.R., van den Heuvel E.P.J.,
664: Habets G.M.H.J., Persi P., 1988, A\&A 198, 200
665:
666: \bibitem{p15} Wex N., Johnston S., Manchester R.N., Lyne A.G., Stappers B.W.,
667: Bailes M., 1998, MNRAS, 298, 997
668:
669:
670:
671:
672:
673: \end{thebibliography}
674: }
675: %\begin{figure}[tb]
676: %\vspace{6cm}
677: %\include{f31.eps}
678: %\hbox{\psfig{file=afig1.eps,height=10cm}}
679: %\caption{The geometry under consideration. P is a pulsar, S is an optical
680: %star, I is a point of interaction of the electrons and photons, O is an
681: %observer. Note that these points are not in one plane in general case.}
682: %\end{figure}
683:
684:
685: %\begin{figure}[t,b]
686: %\centerline{\epsfig{file=f31.eps}}
687: %\vspace*{-10cm}
688: %Figure 1: The geometry under consideration. Point $P$ marks the position
689: %of the pulsar, point $Be$ - the position of the Be star. Curves $S_1$,
690: %$S_2$ represents the shock waves, curve $c$ - the contact surface. Curve
691: %$U$ is a curve of the constant density (see text). Points $A$ and $B$ mark
692: %the places of the curve $U$ intersection with the shocks $S_1$ and $S_2$,
693: %point $C$ is the point of the shock $S_2$ intersection with the binary
694: %axis. Points $D$ and $E$ lies on a little circle with center at $P$ and
695: %on the lines $PC$ and $AP$ correspondingly. \end{figure}
696:
697:
698: %\newpage
699: %\newpage
700: %\centerline{\epsfig{file=f32.eps}}
701: %\vspace*{-13cm}
702:
703: %Figure 2: The spectra of the relativistic particles synchrotron emission,
704: %calculated for the pulsar with an axis along the binary axis at the moment
705: % when it is 30 days to the
706: %the periastron passage. Also here are shown the observational data
707: %that were recieved 3 days before the peristron passage and 16 days after.\\
708: %1.$v_d=c/15, \sigma=5\times 10^{-3}, \th_1=10\degm, \th_2=70\degm$\\
709: %2.$v_d=c/10, \sigma=5\times 10^{-3}, \th_1=10\degm, \th_2=70\degm$\\
710: %3.$v_d=c/10, \sigma=10^{-3}, \th_1=10\degm, \th_2=70\degm$\\
711: %4.$v_d=c/3, \sigma= 10^{-3}, \th_1=20\degm, \th_2=55\degm$\\
712:
713: %\newpage
714:
715: %Figure 3: On this figure the dependence of the degree of polarization on
716: %the angle between the line of sight and the binary axis is shown. It was
717: %calculated for the case of pulsar with an axis along the binary axis and
718: %under the assumption that the effect of the Faraday rotation is
719: %negligible.
720:
721:
722:
723: \end{document}
724:
725:
726:
727:
728:
729:
730:
731:
732:
733:
734:
735:
736: