1: \documentclass[12pt,preprint]{aastex}
2: %\documentstyle[emulateapj,psfig]{article}
3: %\documentclass[preprint2]{aastex}
4:
5: \newcommand{\siml}{\lower4pt \hbox{$\buildrel < \over \sim$}}
6: \newcommand{\simg}{\lower4pt \hbox{$\buildrel > \over \sim$}}
7:
8: \begin{document}
9: \title{A model for the challenging ``bi-drifting'' phenomenon
10: in PSR J0815+09}
11: \author{G. J. Qiao$^{1}$, K. J. Lee$^{1}$, B. Zhang$^{3}$. R. X. Xu$^1$,
12: \& H. G. Wang$^{2}$}
13:
14: \input psfig.sty
15:
16: \affil{ $^1$Department of Astronomy, Peking University,
17: Beijing 100871, China, gjn@pku.edu.cn \\
18: $^2$ Center for Astrophysics, Guangzhou University, Guangzhou
19: 510400, China, cosmic008@263.net\\
20: $^3$ Department of Physics, University of Nevada, Las Vegas,
21: Nevada 89154-4001, bzhang@physics.unlv.edu}
22:
23: \begin{abstract}
24:
25: A new drifting pulsar, PSR J0815+09, was discovered in the Arecibo
26: drift-scan searches \citep{Mc04}. An intriguing feature of this
27: source is that within the four pulse components in the integrated
28: pulse profile, the sub-pulse drifting direction in the two leading
29: components is opposite from that in the two trailing components.
30: In view that the leading theoretical model \citep{RS75} for pulsar
31: sub-pulse drifting can only interpret one-direction sub-pulse
32: drifting, the observed \emph{bi-drifting} phenomenon from PSR
33: J0815+09 poses a great challenge to the pulsar theory. The inner
34: annular gap (IAG), a new type of inner particle accelerator, was
35: recently proposed to explain both $\gamma$-ray and radio emission from
36: pulsars \citep{QLWXH04}. Here we show that the coexistence of the IAG
37: and the conventional inner core gap (ICG) offers a natural
38: interpretation to the bi-drifting phenomenon. In particular, the
39: peculiar drifting behavior in PSR J0815+09 can be reproduced
40: within the inverse Compton scattering (ICS) model for pulsar radio
41: emission.
42:
43: \end{abstract}
44:
45: \keywords{pulsars: general --- pulsars: individual (PSR J0815+09)
46: --- radiation mechanisms: non-thermal --- stars: neutron ---
47: elementary particles}
48:
49: \section{Introduction}
50:
51: The sub-pulse drifting phenomenon observed in many pulsars has been
52: widely regarded as a powerful tool to probe the pulsar inner
53: magnetospheric structure and radiation mechanism (Ruderman \&
54: Sutherland 1975, hereafter RS; Gil, Melikidze \& Geppert 2003).
55: It may be linked to the
56: physical properties of the surface of pulsars (RS75), shedding
57: light on the nature of pulsars, e.g. whether they are normal neutron
58: stars or bare strange stars (Xu, Qiao \& Zhang 1999).
59:
60: In the classical drifting model, sub-pulse patterns are
61: manifestations of some sparks passing along a ring (``carousel'')
62: in the polar cap region circulating the magnetic axis (RS75). This
63: interpretation is supported by detailed analysis of drifting data
64: \citep{DR99,DR01,AD01,GS03}. The RS model is the leading model to
65: interpret the observations quantitatively. More detailed analysis
66: revealed that the circulation speed in a pure vacuum gap is too
67: high when compared with the observations \citep{GGMG03}. Such a
68: discrepancy may be resolved through introducing partial screening
69: of the gap electric field so that the inner gap is not purely
70: vacuum \citep{GGMG03,CR80,UM95}. These models are successuful to
71: interpret many observations, including time variations of the drift
72: rate and changes of the apparent drift direction \citep{GGMG03}.
73:
74: Recently, a new drifting pulsar, PSR J0815+09, was discovered in
75: the Arecibo drift-scan searches \citep{Mc04}. An intriguing
76: feature of this source is that within the four pulse components in
77: the integrated pulse profile, the sub-pulse drifting direction in
78: the two leading components is opposite from the one in the two
79: trailing components. We call this ``bi-direction drifting'' or
80: ``bi-drifting'' phenomenon. This phenomenon poses a great
81: challenge to the RS model and its variants, since in all these
82: models, the drifting direction in different spark-rings is expected to
83: be the same. Recently, Qiao et al. (2004) suggested that an inner
84: annular gap (IAG) likely coexists with the conventional inner core
85: gap (ICG) in pulsars (especially when pulsars are bare strange
86: stars), and proposed a phenomenological model to illustrate the
87: origin of the $\gamma$-ray and radio radiation from pulsars. In
88: this paper, we propose that the bi-drifting phenomenon is a
89: natural consequence of the coexistence of the IAG and the ICG.
90: Furthermore, the complicated drifting patterns in PSR J0815+09
91: could be reproduced if the pulsar radio emission is dominantly
92: generated through coherent inverse Compton scattering (ICS)
93: processes \citep{QL98,XLHQ00,QLZH01}.
94:
95: In \S 2, we introduce both the ICG and the IAG in pulsars, and the ICS
96: model for radio emission is introduced in \S 3. Numerical simulations
97: of the observed bi-drifting phenomenon is presented in \S 4, and the
98: conclusions and discussions are presented in \S 5.
99:
100: \section{The inner core gap (ICG) and the inner annular gap (IAG)}
101:
102: The pulsar's polar region defined by the last open magnetic field
103: lines actually
104: includes two parts separated by the \emph{critical field
105: lines} (RS75, Qiao et al. 2004). Oppositely-charged particles
106: leave the two region respectively, (e.g. the positive charges leave
107: the central region while the negative charges leave the annular region
108: for a pulsar with ${\rm \Omega \cdot B<0}$). The deviation of the
109: charge density from the Goldreich-Julian (1969, hereafter GJ)
110: charge density has opposite signs in the two regions.
111: Qiao et al. (2004) suggested that in principle there could form
112: two kinds of sparking inner gaps, i.e. the
113: conventional inner core gap (ICG) above the central part of the
114: polar region and an inner annular gap (IAG) above the annular part of
115: the polar region. Both gaps can be high energy particle accelerators.
116: A geometric model involving
117: emission from both gaps can interpret the diverse morphology
118: of both gamma-ray and radio emission from gamma-ray pulsars. The
119: geometric structure of the two kinds of polar gap is plotted in
120: Fig.\ref{cap}.
121:
122: \begin{figure}
123: \plotone{f1.eps} \caption{The inner annular gap (IAG) and the
124: inner core gap (ICG). $r_{\rm p, IAG}$ and $r_{\rm p, ICG}$ are
125: the radii of the IAG and the ICG, respectively. The two
126: gaps are divided by the boundary line 'ch', which is the critical
127: magnetic line that passes through the intersection of the null-charge
128: line and the light cylinder. The dashed lines are used to
129: illustrate the electric property of the gaps discussed in section
130: 2. The parameters used in our simulations are presented in the text.
131: }
132: \label{cap}
133: \end{figure}
134:
135: For the inner gaps, the typical gap height is
136: smaller than the polar cap radius, and one can use the 1-D
137: approximation of the Poisson equation in the co-rotate frame,
138: i.e. $\partial E_{\parallel}/\partial
139: x=4\pi(\rho-\rho_{\rm GJ})$, where $x$ is the longitudinal
140: distance measured from the surface along the curved magnetic field
141: lines, and $E_{\parallel}$ is the parallel component of the
142: electric field with respect to the magnetic field, $\rho$ is the charge
143: density and $\rho_{\rm GJ}$ is the GJ charge density. This
144: equation governs the electric field along the magnetic field
145: (i.e. segments ``ab'', ``ch'', ``dg'' and ``ef'' in Fig.\ref{cap}).
146: The boundary condition equation is $E_{\parallel}=0$ at the upper
147: boundary of the gap (i.e. segment ``edcb'').
148:
149: When the gaps are re-generated after each sparking process, since
150: opposite charged particles leave the ICG and the IAG regions,
151: respectively, the sign of $\rho-\rho_{\rm GJ}$ is the opposite in
152: the two gaps. The Poisson equation above shows that the directions
153: of $E_{\parallel}$ are also different in the IAG and ICG.
154: We expect that at the boundary between ICG and IAG (line ``ch'')
155: the parallel electric field vanishes. That is $\rm \int_{c}^{h}
156: E\cdot ds=0$.
157:
158: We handle the electrodynamics in the co-rotate frame. For any
159: close circuits (e.g. ``defg'', ``abch''), one has $\oint E\cdot
160: ds=0$. In the closed magnetic field region ${\bf E \cdot B}=0$, so
161: $\int_{e}^{f}{\bf E\cdot ds}=0$. As discussed in RS75, we have
162: $\int_{f}^{g}{\bf E\cdot ds}=0$. Thus for the IAG, one has
163: $\int_{g}^{d} {\bf E\cdot ds}+\int_{d}^{e} {\bf E\cdot ds}=0$.
164: When considering the boundary condition between the two gaps one has
165: $\int_{c}^{h}{\bf E\cdot ds}=0$. Also $\int_{h}^{a}{\bf E\cdot
166: ds}=0$ is satisfied for the same reason as in the discussion of the IAG
167: case. One can get $\int_{a}^{b} {\bf E\cdot ds}+\int_{b}^{c}
168: {\bf E\cdot ds}=0$. Therefore for the IAG and the ICG the
169: perpendicular electric field is directly linked to the parallel
170: electric field.
171: Becasue the parallel electric fields in the IAG and the ICG have
172: different directions (i.e. $\int_{a}^{b} {\bf E\cdot ds}$ and
173: $\int_{g}^{d} {\bf E\cdot ds}$ have different signs), the
174: perpendicular
175: electric fields in the two gaps are also different. The drifting
176: velocity ${\bf v=E \times B/|B|^2}$ have opposite signs in the ICG
177: and the IAG, because the direction of magnetic fields is the same
178: in the two gaps. This proposes a fundamental physical process to
179: understand the bi-drifting phenomenon.
180:
181: \section{The inverse Compton scattering (ICS) model of pulsar radio
182: emission}
183:
184: Any radiation model possessing a symmetrical radiation beam
185: is not applicable to PSR J0815+09. If the radiation beam
186: is symmetrical, the 2nd and 3rd components should come from the same
187: beam, so they should have a same drifting sense which contradicts the
188: data. When the aberration and retardation effects are considered, the
189: ICS (inverse Compton scattering) model naturally predicts an asymmetric
190: beam and offers a natural mechanism to interpret the phenomenon.
191:
192: The ICS model \citep{QL98,XLHQ00,QLZH01,QLWXH04} suggests that the
193: observed radio emission comes from different emission heights and each
194: sparking ring can, in principle, lead to several emission cone-shaped
195: beams. Following Qiao and Lin (1998), one can get the so-called
196: ``beam-frequency figure'' of the ICS model, which naturally gives rise
197: to a narrow central core emission component and two conal components,
198: meeting Rankin's (1983; 1993) empirical proposal from radio pulsar
199: data. Another remark is that the three-component scheme is an average
200: picture, which resembles Manchester's (1995) ``window function''
201: scheme. Depending on the line of sight, different number of emission
202: components at a particular observing frequency can be observed. That
203: PSR J0815+09 is observed as a drifting pulsar means that the line
204: of sight sweeps across the beam rim rather than cutting the rim
205: tangentially. For the latter case periodical variation of the
206: pulse intensity rather than sub-pulse drifting should be observed. It
207: is likely that the line of sight misses the narrow core beam and only
208: sweeps the two conal beams. In this way, there are four conal
209: radiation beams in the system, two of which come form the IAG while the
210: other two come from the ICG. Generally the four beams should form
211: eight pulse components. There are still several important issues for
212: determining the beam morphology. One thing to put in consideration is
213: the aberration and retardation effects. Another one is the radiation
214: process of inverse Compton scattering.
215:
216: Because different emission components are emitted at different
217: heights, the retardation and aberration effects must be taken into
218: account, and they are found to be important to determine the real
219: emission morphology. These two effects will smear the leading
220: radiation components and strengthen the trailing components (Qiao \&
221: Lin 1998, Qiao et al. 2004, Dyks et al. 2004). From the simulations
222: (Qiao \& Lin 1998, Dyks et al. 2004), it is found that the trailing
223: components are about 2 orders of magnitude brighter than the leading
224: components if the radio waves are superposed coherently.
225: For incoherent superposition, the contrast should be still 1 order of
226: magnitude.
227:
228: In the ICS model, the observed radiation comes from the inverse Compton
229: scattering between the high energy secondary particles and the low
230: frequency electromagnetic wave (EM wave) generated by the polar gap
231: sparking process. The intensity of the low frequency EM wave is a
232: function of the direction angle $\chi$ (i.e. $\propto \sin^{2}\chi$),
233: where $\chi$ is the angle between the radiation direction and the
234: electric dipole moment. When taking the rotation effect into account,
235: the leading and the trailing radiation sites at a same height do not
236: have the same angle respected to the direction of the spark's electric
237: moment, so that they do not possess the same intensity of the low
238: frequency EM wave. This effect will intensify the leading components
239: and weaken the trailing components. The ratio of the leading low
240: frequency wave intensity ($I_{\rm L}$) to trailing intensity ($I_{\rm
241: T}$) can be given by the equation $I_{\rm L}/I_{\rm
242: T}=\sin^{2}(\theta+\sin^{2}\theta)/\sin^{2}(\theta-\sin^{2}\theta)$,
243: when the inclination angle is $90^{o}$ and the spark is a pure
244: dipole. This effect can lead to an intensity contrast of 2
245: orders of magnitude between the leading and trailing pulse components
246: for coherent radiation or of 1 order of magnitude for incoherent
247: radiation. Beside the two effects above, other potential
248: mechanisms may also play a role. For example, if the polar gap
249: sparks are triggered by the incoming $\gamma$-ray photons, the geometry
250: configuration of the pulsar, the beam direction and the $\gamma$-ray
251: photon source direction respected to the pulsar will all affect the
252: intensity ratio of the leading and the trailing components.
253:
254: Comparing among these effects, the first one is the common situation.
255: So in most cases, the trailing components may be the ones that are
256: observed. However, we can not exclude other possibilities discussed
257: above, and will take it as a basic assumption that the radiation beam
258: is asymmetrical, and that only one half of the components can be
259: observed. Such an assumption has received observational
260: supports, since some pulsars are already observed to have asymmetrical
261: pulse profiles. This assumption can be tested by high quality
262: polarization observations, which may indicate whether the leading or
263: the trailing components are observed.
264:
265: \section{Simulations of sub-pulse drifting}
266:
267: To simulate the drifting pulse patterns, two parameters are included,
268: i.e. the number of the sub-beams in each beam and the drifting rate of
269: each pulse component. The numbers of the sub-beams are calculated
270: theoretically, and the drifting rates can be obtained from the
271: simulations, which would be used to infer the dynamical
272: structure of both gaps.
273:
274: Gil \& Sendyk (2000) suggests that
275: the number of the sparks that can be observed
276: in drifting pattern is given by $N \simeq 2 \pi r_{\rm D}/h$,
277: where $N$ is the number of sparks, $h$ is the height of the
278: gap, $r_{\rm D}$ is the radius of the sub-pulse drifting trajactory
279: which is given by $r_{\rm D}=r_{\rm p}-h/2$. For the parameters of PSR
280: J0815+09 (P=0.645s and $B=3\times10^{11}$G, McLaughlin et al.
281: 2004), it can be shown that $r_{\rm p,\rm IAG}=R^{3/2} R_{\rm
282: lc}^{-1/2}\simeq (1.45 \times 10^4 {\rm cm}) p^{-1/2}\simeq 1.80 \times
283: 10^4 $ cm for the IAG and $r_{\rm p, ICG} =(2/3)^{3/4}R^{3/2}R_{\rm
284: lc}^{-1/2}\simeq(1.07 \times 10^4 {\rm cm} )p^{-1/2}\simeq 1.35 \times
285: 10^4$ cm for the ICG, where $R$ is the stellar radius, $R_{\rm lc}$ is
286: the radius of light cylinder. For a resonance inverse Compton
287: scattering gap, the gap height is $h\simeq (1.1\times 10^3 {\rm cm})
288: B_{12}^{-1}p^{1/3} \simeq 3.2\times 10^3 $ cm \citep{ZQH97}, where
289: $B_{12}$ is the surface magnetic field in unit of $10^{12}$G, where
290: the curvature radius of the magnetic fields is taken as $10^6 $cm. So
291: the IAG holds $N \simeq 33$ observed sparks while the ICG holds $ N
292: \simeq 21$ sparks. These two parameters will be put in our
293: simulation for the drifting patterns. Assuming a same intensity for
294: each component within the gap, we can also calculate
295: the integrated pulse profile.
296:
297: The drifting rates of the sparks depend on the electric structure of
298: the gaps. However many uncertainties involved in the gaps
299: \citep{ZQLH97,GGMG03,QLWXH04} prevent us from retrieving solid
300: information for the drifting rates. Here we use the observational
301: drifting rates to constrain the gap models. From the theoretical
302: sub-beam number and the observed drifting patterns, we get the drifting
303: rates in both gaps. The simulated results and observations are
304: compared in Fig.\ref{drift}.
305:
306: \begin{figure}
307: \plotone{f2.eps}
308: \caption{The observed sub-pulse drifting patterns and pulse profiles
309: as compared with our
310: simulations. The left two panels are the patterns of
311: drifting sub-pulses. The right two panels are the integrated pulse
312: profiles. The observational data are derived from McLaughlin et
313: al. 2004. In the simulation, the ICG has 21 sparks circularly drifting
314: around the magnetic axis with a period of 210s, while the IAG has 33
315: sparks with a drifting period of 310s. The spark number is given by
316: theoretical estimates and the drifting periods are inferred through
317: matching the simulation with the observations.
318: Lacking of polarization data, we can not constrain the radiation
319: geometry. In the simulation, the radiation geometry are choosen
320: relatively arbitrarily. The angular radii for the four beams are 12,
321: 28, 46 and 63 degrees, the inclination angle is 45 degrees, and the
322: impact angle is 5 degrees.
323: }
324: \label{drift}
325: \end{figure}
326:
327: \section{Results and discussion}
328: By considering both the ICS process and the sparks from the IAG
329: and the ICG, we have demonstrated that the bi-drifting phenomenon
330: observed in PSR J0815+09 is naturally interpreted. This result also
331: lends support to the existence of the IAG. Some parameters
332: and effects related to our simulation are discussed as follows.
333:
334: (1) The gap heights. Assuming that the spark diameter in the polar
335: gap is the same as the gap height \citep{GS00}, we have used the
336: resonance inverse Compton scattering induced polar gap model
337: \citep{ZQLH97} to calculate the gap height to get the spark number in
338: the polar cap. One requirement to the height of the sparks is that
339: some sparks must take place in the IAG. The height of a resonance
340: inverse Compton scattering induced gap in our simulation is consistent
341: with this condition.
342:
343: (2) The drifting rates. The drifting rates derived from the
344: observation give us insights into the electric structure of the two
345: gaps. The average parallel electric field $E_{\parallel}$ in the gap
346: can be estimated with $E_{\parallel}\sim 4\pi h(\rho-\rho_{\rm GJ})$.
347: From \S 2, we know that the perpendicular electric field
348: $E_{\perp}$ can be estimated as $E_{\rm \perp}\simeq
349: E_{\parallel}h/\Delta r$ (see also Gil et al. 2003), where
350: $\Delta r$ is the transverse spatial dimension of the sparks and
351: has $\Delta r\simeq h/2\simeq1.6 \times 10^3 $ cm. The drifting
352: velocity of the sparks is therefore $v_{\rm D}=E_{\perp} c/B$,
353: where $c$ is the speed of light, and the period for a spark to make a
354: circle in both gaps can be written as
355: $\hat{P_{3}}=2\pi r_{\rm D}/v_{\rm D}$, or
356: \begin{equation}
357: \hat{P_{3}}\simeq \frac{B r_{\rm D} \Delta r} {2ch^{2}(\rho-\rho_{\rm GJ})}
358: \label{P3}
359: \end{equation}
360: The period for a subpulse to reappear at a same phase is defined by
361: $P_3 = \hat P_3/n$, where $n \propto r_{\rm D}$ is the number of
362: sparks in a circle.
363: Since both the IAG and the ICG have the same height $h$ and hence, the
364: same $\Delta r$, from Eq.(\ref{P3}) one gets
365: \begin{equation}
366: \frac{P_{\rm 3,ICG}}{P_{\rm 3,IAG}} \simeq \frac { r_{\rm \rm D,
367: IAG} \hat{P_{\rm 3}}_{\rm , ICG} } { r_{\rm D, ICG} \hat{P_{\rm
368: 3}}_{\rm , IAG} } \simeq \frac{(\rho-\rho_{\rm GJ})_{\rm
369: IAG}}{(\rho-\rho_{\rm GJ})_{\rm ICG}} =
370: \frac{E_{\parallel, \rm
371: IAG}}{E_{\parallel, \rm ICG}}, \label{ele}
372: \end{equation}
373:
374: The simulation gives $\hat{P_{\rm 3,}} _{\rm IAG}/\hat{P_{\rm
375: 3,}}_{\rm ICG}\simeq1.5$, and from \S 4, we have $r_{\rm D,
376: IAG}/r_{\rm D, ICG}\simeq1.4$ (Fig.~\ref{cap}). So
377: $E_{\parallel, \rm IAG}/E_{\parallel, \rm ICG}=P_{3,\rm IAG} /P_{3,\rm
378: ICG} \simeq0.9$.
379: This indicates that the absolute charge density deviation from
380: $\rho_{\rm GJ}$ in the ICG is roughly the same as that in the IAG,
381: although a charge deficit presents in the ICG while a charge excess
382: presents in the IAG. Also the $E_{\parallel}$ in the ICG is roughly
383: the same as in the IAG but with a different sign. The sub-pulse
384: drifting phenomenon offers a diagnostic tool for the plasma near the
385: polar cap region.
386:
387: The theoretical drifting period of a vacuum gap for PSR J0815+09
388: should be order of 1 s (Eq. (\ref{P3}) and put $\rho=0$) and is
389: much shorter than the fitted value $\sim10^2$ s. This has been
390: also noticed earlier by Gil et al. (2003). There are three
391: possible ways to solve this problem. One is to conclude that both
392: the IAG and the ICG are not vacuum gaps, so that only a small
393: charge density deviation from the GJ charge density is allowed. In
394: order to match the observations, only a $\siml 1\%$ deviation from
395: the GJ density in both gaps is required. This would naturally
396: reduce the drifting rates, and is consistent with the partial
397: screening picture (Gil et al. 2003). It is also consistent with
398: the XMM-Newton observation results for another drifting pulsar PSR
399: 0943+10 (B. Zhang, D. Sanwal \& G. G. Pavlov 2004, in
400: preparation). In such a case, an unsteady space charge limited
401: flow may play an important role, and a detailed model is called
402: for. The second reason is overestimating the drifting
403: velocity, because the method above just give the average value of
404: $E_{\perp} $ which is larger than that at sparking location. The
405: third way to solve the problem is to take into account the
406: drifting dynamics, in which the drifting velocity is not the
407: $E\times B$ velocity. Again more detailed investigations are
408: needed. It should be noted that the subpulse drifting in the two
409: gaps depends on the global electric property of the pulsar. Two
410: gaps interact with each other via the boundary condition. So a
411: physically reasonable and complete drifting subpulse model should
412: at least includes a global electric solution of pulsars.
413:
414: It is found that within the observation data span, the four pulse
415: components keep a phase relationship at least within 120 pulses
416: (77s). Our Eq.(2) suggests that $P_3$ in both gaps are roughly the
417: same. A further physical possibility is that the two gaps may interact
418: with each other. Dynamical system theories prove that for two
419: quasi-periodical systems, if there are some small nonlinear
420: interactions, the two systems will be locked into a nearby
421: frequency with a rational ratio \citep{JBB83}. Because of the
422: interactions between the two gaps, the two drifting
423: frequencies may be locked into a relative constant ratio in our
424: model, so that the apparent $P_3$ could appear the same for all four
425: pulse components. Detailed interaction dynamics is needed to further
426: address this problem. Some other kinds of interaction between the two
427: gaps are also possible (e.g. Young 2004).
428:
429: (3) High quality polarization observations are needed to finally
430: verify whether the observed components are the trailing or the leading
431: components.
432:
433: In summary, the so called \emph{bi-drifting} phenomenon is a newly
434: observed touch-stone to test the radiation theories and the surface
435: physical properties of the pulsar. Our simulated results support the
436: coexistence of IAG and ICG. Further investigations are needed to
437: address the questions such as how such gaps are generated.
438:
439: \begin{acknowledgements}
440: We are very grateful to the referee's insightful comments, which led
441: us to improve some of the very important points raised in this
442: paper. We also thank Drs. Esamdin Ali \& Han, J. L. for their valuable
443: discussions. This work is supported by NSF of China (10373002,
444: 10273001). B.Z. acknowledges NASA NNG04GD51G for support.
445:
446:
447: \end{acknowledgements}
448:
449: \begin{thebibliography}{}
450:
451: \bibitem[Asgekar \& Deshpande 2001]
452: {AD01}Asgekar, A., \& Deshpande, A. 2001, \mnras, 326, 1249
453:
454: \bibitem[Cheng \& Ruderman 1980]
455: {CR80} Cheng, A. F., \& Ruderman, M. A. 1980, \apj, 235, 576
456:
457: \bibitem[Deshpande \& Rankin 1999]{DR99}
458: Deshpande, A. A., \& Rankin, J. M. 1999, \apj, 524, 1008
459:
460: \bibitem[Deshpande \& Rankin 2001]{DR01}
461: Deshpande, A. A., \& Rankin, J. M. 2001, \mnras, 322, 438
462:
463: \bibitem[Dyks et al. 2004]{DRH04}
464: Dyks, J., Rudak, B. \& Harding, A. K. 2004, \apj, 607, 939
465:
466: \bibitem[Gil \& Sendyk 2000]{GS00}
467: Gil, J. A., \& Sendyk, M., 2000, \apj, 541, 351
468:
469: \bibitem[Gil et al. 2003]{GGMG03}
470: Gil, J. A. ,Melikidze, G. I. \& Geppert, U., 2003, \aap, 407, 315 (astro-ph/0305463)
471:
472: \bibitem[Gil \& Sendyk 2003]{GS03}
473: Gil, J. A., \& Sendyk, M. 2003, \apj, 585, 453
474:
475: \bibitem[Goldreich \& Julian 1969]{GJ69}
476: Goldreich, P. \& Julian, W. H., 1969, \apj, 157, 869
477:
478: \bibitem[Jensen et al. 1983]{JBB83}
479: Jensen, M. H., Bak, P., Bohr, T. \prl, 50, 1637
480:
481: \bibitem[Manchester 1995]{M95}
482: Manchester, R. N. 1995, J. Astrophys. Astr., 16, 107
483:
484: \bibitem[McLaughlin et al. 2004]{Mc04}
485: McLaughlin, M. A., Lorimer, D. R., Champion, D. J., et al. 2004, in Young Neutron Stars
486: and Their Environments IAU Symposium, Vol. 218 (eds. F. Camilo and
487: B. M. Gaensler) (astro-ph/0310454)
488:
489: \bibitem[Qiao et al. 2004]{QLWXH04}
490: Qiao, G. J., Lee, K. J., Wang, H. G., Xu, R. X., \& Han, J. L., 2004, \apj, 606, L49
491:
492: \bibitem[Qiao \& Lin 1998]{QL98}
493: Qiao, G. J.,\& Lin, W. P., 1998, \aap, 333, 172
494:
495: \bibitem[Qiao et al. 2001]{QLZH01}
496: Qiao, G. J., Liu, J. F., Zhang, B. \& Han, J. L. 2001, \aap, 377, 964
497:
498: \bibitem[Rankin 1983]{R83}
499: Rankin, J. M. 1983, \apj, 274, 333
500:
501: \bibitem[Rankin 1993]{R93}
502: Rankin, J. M. 1993, \apjs, 85, 145
503:
504: \bibitem[Ruderman \& Sutherland 1975]{RS75}
505: Ruderman, M. A., \& Sutherland, P. G., 1975, \apj, 196, 51
506:
507: \bibitem[Usov \& Melrose 1995]{UM95}
508: Usov, V. V., \& Melrose, D. B. 1995, Austr. J. Phys., 48, 571
509:
510: \bibitem[Xu et al. 2000]{XLHQ00}
511: Xu, R. X., Liu, J. F., Han, J. L., \& Qiao, G. J., 2000, \apj, 535, 354
512:
513: \bibitem[Xu et al. 1999]{XQZ99}
514: Xu, R. X., Qiao, G. J. \& Zhang, B. 1999, \apj, 522, L109
515:
516: \bibitem[Young 2004]{Young04}
517: Young, M. D. T. 2004, in Young Neutron Stars and Their Environments IAU Symposium,
518: Vol. 218 (eds. F. Camilo and B. M. Gaensler) (astro-ph/0310411)
519:
520: \bibitem[Zhang et al. 1997a]{ZQH97}
521: Zhang, B., Qiao, G.J., Han, J.L. 1997, \apj, 491, 891
522:
523: \bibitem[Zhang et al. 1997b]{ZQLH97}
524: Zhang, B., Qiao, G. J., Lin, W. P. \& Han, J. L. 1997, \apj, 478, 313
525:
526: \end{thebibliography}
527:
528:
529: \end{document}
530: