1: %\documentclass{emulateapj}
2: \documentclass[12pt,preprint]{aastex}
3: %\documentclass[referee]{aastex}
4: \usepackage{emulateapj5}
5: \usepackage{graphicx}
6: \usepackage{natbib}
7:
8: \slugcomment{ApJ accepted}
9:
10: \begin{document}
11:
12: \title{Non-Thermal X-ray Properties of Rotation Powered Pulsars and Their Wind Nebulae}
13:
14: \author{Xiang-Hua Li\altaffilmark{1,2}, Fang-Jun Lu\altaffilmark{1} and Zhuo Li\altaffilmark{3}}
15:
16: \altaffiltext{1}{Laboratory of Particle Astrophysics, Institute of High Energy Physics,
17: CAS, Beijing 100049, China; lixh@ihep.ac.cn; lufj@ihep.ac.cn}
18: \altaffiltext{2}{Max-Planck-Institut f\"{u}r Radioastronomie,
19: Auf dem H\"ugel 69, 53121 Bonn, Germany}
20: \altaffiltext{3}{Physics Faculty, Weizmann Institute of Science, Rehovot 76100, Israel;
21: lizhuo@weizmann.ac.il}
22: \date{Received / Accepted}
23:
24: \begin{abstract}
25: We present a statistical study of the non-thermal X-ray emission of 27 young rotation
26: powered pulsars (RPPs) and 24 pulsar wind nebulae (PWNe) by using the {\sl Chandra} and
27: the {\sl XMM-Newton} observations, which with the high spatial resolutions enable us to
28: spatially resolve pulsars from their surrounding PWNe. We obtain the X-ray luminosities
29: and spectra separately for RPPs and PWNe, and then investigate their distribution and
30: relation to each other as well as the relation with the pulsar rotational parameters. In
31: the pair-correlation analysis we find that: (1) the X-ray (2-10 keV) luminosities of both
32: pulsar and PWN ($L_{\rm X,psr}$ and $L_{\rm X,pwn}$) display a strong correlation with
33: pulsar spin down power $\dot{E}$ and characteristic age $\tau$, and the scalings
34: resulting from a simple linear fit to the data are $L_{\rm
35: X,psr}\propto\dot{E}^{0.92\pm0.04}$ and $L_{\rm X,pwn}\propto\dot{E}^{1.45\pm0.08}$ (68\%
36: confidence level), respectively, however, both the fits are not statistically acceptable;
37: (2) $L_{\rm X,psr}$ also shows a possible weak correlation with pulsar period $P$ and
38: period derivative $\dot{P}$, whereas $L_{\rm X,pwn}$ manifests a similar weak correlation
39: with $\dot{P}$ only; (3) The PWN photon index $\Gamma_{\rm pwn}$ is positively correlated
40: with $L_{\rm X,pwn}$ and $L_{\rm X,pwn}/\dot{E}$. We also found that the PWN X-ray
41: luminosity is typically 1 to 10 times larger than that from the underlying pulsar, and
42: the PWN photon indices span a range of $1.5\lesssim\Gamma_{\rm pwn}\lesssim2$. The
43: statistic study of PWN spectral properties supports the particle wind model in which the
44: X-ray emitting electrons are accelerated by the termination shock of the wind.
45: \end{abstract}
46:
47: \keywords{radiation mechanisms: non-thermal - star: neutron - stars: pulsar: general - X-rays: general}
48:
49:
50: \section{Introduction}
51: The rotation-powered pulsars (RPPs) are known as rapidly spinning and strongly magnetized
52: neutron stars that are radiating at the expense of their rotational energy. The X-ray
53: emission of RPPs may contain both thermal and non-thermal components. The thermal emission
54: might be further divided into non-pulsed and pulsed components. The non-pulsed component,
55: originates from the cooling of the neutron star,
56: is from the whole pulsar surface with a characteristic temperature of about 0.1 keV,
57: while the pulsed component comes from hot spots
58: ($\sim$ 1.0 keV) on the pulsar surface, which are heated by the bombardment of
59: relativistic particles streaming back to the surface from the pulsar magnetosphere (Kundt
60: \& Schaaf 1993, Zavlin et al. 1995, Gil \& Krawczyk 1996). The non-thermal pulsar emission
61: is from the pulsar magnetosphere, and it might also contain pulsed (e.g., Cheng \& Zhang
62: 1999; Zhang \& Harding 2000) and non-pulsed components (e.g., Tennant et al. 2001; Becker
63: et al. 2004). In some cases, a pulsar wind nebula (PWN) is found to surround a RPP. The
64: X-ray emission of the PWN is non-pulsed and often dominates the non-thermal emission of
65: the system.
66:
67: A lot of efforts have been devoted to the statistical studies of pulsar X-ray emission
68: properties, with particular emphasis on the efficiency of conversion of the pulsar spin
69: down power $\dot{E}$ into X-ray luminosity. By using the {\sl Einstein} data, Seward \&
70: Wang (1988) found that $L_{\rm X}\propto \dot{E}^{1.39}$, where $L_{\rm X}$ is the
71: 0.2-4.0 keV X-ray luminosity of the pulsar (plus PWN). Becker \& Tr{\"u}mper (1997)
72: obtained $L_{\rm X}\simeq 10^{-3}\dot{E}$ using a sample of 27 pulsars observed by {\sl
73: ROSAT}, where $L_{\rm X}$ is the total X-ray (0.1-2.4 keV) luminosity of the pulsar plus
74: PWN. However, in these two works, the thermal emission may contribute significantly to
75: the total X-ray luminosity given the adopted soft X-ray band. Saito (1998) analyzed 16
76: RPPs observed by {\sl ASCA} (2-10 keV) and found $L_{\rm X}\propto \dot{E}^{3/2}$.
77: Possenti et al. (2002) reported $L_{\rm X}\propto \dot{E}^{1.34}$ using 39 pulsars
78: observed by {\sl ASCA, RXTE, BeppoSAX, Chandra} and {\sl XMM-Newton}. The X-ray
79: luminosities in Saito (1998) and Possenti et al. (2002) also include the total emission
80: due to the pulsars plus PWNe, given the limited spatial resolutions of {\sl ASCA}, {\sl
81: RXTE}, and {\sl BeppoSAX}. Cheng et al. (2004) divided the total X-ray emission into
82: pulsed and non-pulsed components, and found that the X-ray luminosity of the pulsed one
83: follows $L_{\rm X,pul}\propto\dot{E}^{1.2}$, which agrees with the model prediction
84: $L_{\rm X}\propto\dot{E}^{1.15}$ by Cheng \& Zhang (1999). For non-pulsed emission, they
85: got $L_{\rm X,npul}\propto\dot{E}^{1.4}$, where they supposed that the non-pulsed
86: component comes mainly from PWNe and the contribution of non-pulsed component from
87: pulsars is negligible. It is worth noting that the scatter in the relation is large,
88: as pointed out by Possenti et al. (2002), who performed study including the estimates of the observational
89: errors and showing that the linear fit is statistically unacceptable.
90:
91: All these previous works suffer from the low spatial resolution of the detectors, making
92: it difficult to resolve the emission of the pulsars from that of their surrounding PWNe.
93: It is the purpose of our current work to resolve and to analyze the pulsar and the PWN emission
94: separately. Thanks to the high spatial resolution observations performed with {\sl
95: Chandra} and {\sl XMM-Newton}, we have been able to satisfactorily investigate 27 pulsars
96: and 24 PWNe, for which we have determined the non-thermal X-ray fluxes and spectra in
97: the 2-10 keV band. Then we have carried out separated statistic studies of RPPs and PWNe,
98: and tested the consistence of their emission properties with current models. The
99: organization of this paper is as following: the sample and the data processing are presented
100: in section 2; the statistical analyses of the X-ray spectral properties of RPPs and PWNe are
101: given in section 3; we discuss the physical implications of our results in section 4 and
102: summarize our work in section 5.
103:
104: \section{Sample and Data processing}
105: We collect pulsar and PWN samples from the observations by {\sl Chandra} and {\sl
106: XMM-Newton}, which both have high spatial resolutions, i.e., $\sim1\arcsec$ and
107: $\sim6\arcsec$, respectively. We take the {\sl Chandra} data directly from the
108: literatures, and if there are no published results, we analyzed the data in this paper.
109: The {\sl XMM-Newton} data are adapted only if there are no relevant data from {\sl
110: Chandra} for the same source. All the {\sl XMM-Newton} results are taken from
111: literatures. Totally we obtain the X-ray spectra of 27 RPPs and 24 PWNe. In our samples,
112: millisecond pulsars (MSPs) are not included. It is generally believed that MSPs have ever
113: undergone an accretion-driven spin-up phase and they are usually old and regarded as a
114: significantly different class. Similar study on the MSPs is also limited by the rare data
115: available. Therefore we do not analyze MSPs here, although we discuss them when compare
116: our analysis with the previous work including MSPs.
117:
118: In our samples, there are 15, out of 27, spectra of pulsars obtained from the
119: archived {\sl Chandra} data. We select only the pulsars detected by the Advanced CCD
120: Imaging Spectrometer (ACIS) in the Timed Exposure (TE) Mode, in which a pulsar is able
121: to be resolved spatially from its surrounding PWN. We calibrate the data with CIAO
122: (ver 3.4) and CALDB (ver 3.3.0). We first reprocess the Level 1 data
123: for the correction of the charge transfer inefficiency (CTI) effects, then
124: clean the background and remove the afterglow. Time intervals
125: with anomalous background rates associated with particle flare events are further
126: rejected in the Level 2 data. And then the pulsar positions are obtained by the {\it
127: celldetect} tool in CIAO. Finally, the spectra are extracted from the Level 2 data and then fit
128: with {\sl XSPEC}. We use both the power-law (PL) and the power-law+blackbody (PL+BB)
129: models to fit the pulsar spectra. If the resulted spectral indices are consistent
130: within errors in both models, then the results from the PL model are used, otherwise
131: those from the PL+BB model are used. In our spectral analysis, we show errors at the
132: 90\% confidence level.
133:
134: Pileup occurs when more than one photon are collected in one pixel within a CCD
135: readout frame, since those photons can only be recorded as a single photon event whose
136: energy is the sum of the collected photons. Therefore pileup may affect the
137: results of spectral analysis. According to section 6.14.2 in the {\sl Chandra} Proposers'
138: Observatory Guide v.7\footnote{http://cxc.harvard.edu/proposer/POG/}, the effect of
139: pileup can be omitted if the pileup fraction is $\le$10\%. However, pileup does affect
140: the spectral analysis even if the pileup fraction is $\le$10\%. For example, since the
141: pileup fraction of PSR J1930+1852 is estimated to be only 6\%, its spectral index is
142: reported to be $1.09^{+0.08}_{-0.09}$ without pileup correction (Lu et al. 2002),
143: whileas the spectral index is $1.35^{+0.06}_{-0.10}$ after pileup correction (Camilo
144: et al. 2002). In our spectral analysis, we first estimate the pileup fraction using
145: PIMMS\footnote{http://cxc.harvard.edu/toolkit/pimms.jsp}, and then add a pileup model
146: in the spectral fitting of the pulsars if the pileup fraction is higher than 3\%.
147: Totally, there are 8 pulsars in which the pileup model is included in the spectral
148: fitting, i.e., PSRs J0205+6449, J0537$-$6910, B0540$-$69 (and its PWN), B0833$-$45
149: (Vela), J1747$-$2958, J1846$-$0258, J1930+1852 and B1951+32.
150:
151: The absorption column density ($N_{\rm H}$) is obtained in several ways. (1) For 6
152: pulsars (PSRs J0205+6449, J0537$-$6910, J1747$-$2958, J1846-0258, J1930+1852 and
153: B1951+32) with bright PWNe, $N_{\rm H}$ are obtained from the spectral fitting of
154: their PWNe, and then fixed when fitting the spectra of the pulsars. (2)
155: PSRs B0540$-$69 and J1124$-$5916 are embedded in SNRs 0540$-$69.3 and G292.0+1.8,
156: respectively, and their PWN spectra below 2.5 keV are strongly affected by the SNRs.
157: At the same time, constraining $N_{\rm H}$ with emission above 2.5 keV is difficult
158: because of the small absorption in this high energy range. Therefore their $N_{\rm H}$
159: are obtained by fitting the pulsar spectra that the contamination of the SNR emission
160: is negligible, and then $N_{\rm H}$ are fixed in the spectral fitting of their PWNe.
161: (3) The PWNe associated with PSRs B0355+54, J1617$-$5055, B1823$-$13, B1929+10 and
162: J2229+6114 are not bright, and $N_{\rm H}$ is determined by jointly fitting the
163: spectra of both the pulsar and its PWN. (4) PSRs J0633+1746 and B0833$-$45 have been
164: studied extensively, and their $N_{\rm H}$ values used in our spectral fitting are
165: taken from Caraveo et al. (2004) and Pavlov et al. (2001).
166:
167: The luminosity uncertainty is crucial in our analysis of correlations, and
168: should be considered carefully. Since the X-ray luminosity is given by $L_{\rm X}=4\pi
169: d^2f_{\rm X}$, where $d$ is the pulsar distance and $f_{\rm X}$ is the 2-10 keV X-ray
170: flux, the $L_{\rm X}$ uncertainty should be derived from the uncertainties of both
171: $f_{\rm X}$ and $d$. The uncertainty of $f_{\rm X}$ is derived from those of the
172: normalization and the photon index in the spectral fitting. For the fluxes taken from
173: literatures, their uncertainties are extrapolated from the published ones by the
174: ratios of the fluxes in 2-10 keV to those in the corresponding published energy ranges.
175:
176: The distances are usually not well constrained, thus the distance uncertainty
177: may dominate the luminosity uncertainty. There are several cases in our
178: samples: (1) the distances of 7 pulsars are derived from the radio dispersion
179: measures, and their errors are conservatively taken to be 40\%, as estimated by Cordes
180: \& Lazio (2001); (2) the distances of 14 pulsars are obtained via
181: their associated SNRs, and some of them are shown with published distance errors in
182: literatures, while for the others without published errors a conservative error of
183: 50\% is taken; (3) PSRs J0537$-$6910 and B0540$-$69 are both located in the Large
184: Magellanic Cloud (LMC), whose distance is taken as 50 kpc, and an
185: error of 10 kpc for a conservative estimation is adapted (Bradley 2007).
186:
187: We summarize the properties of all the 27 RPPs and 24 PWNe in Tables 1 and
188: 2, respectively. In Table 2 only 22 PWNe are associated with the RPPs listed in
189: Table 1. The other two pulsars are excluded from Table 1: (1) Camilo et al. (2004)
190: suggested that PSR J1016$-$5857 is too faint to be resolved from its background PWN in
191: the {\it Chandra} Observation; (2) Hessels et al. (2004) found that the spectrum of
192: PSR J2021+3651 can be fit by a BB model and is thus dominated by thermal components.
193: On the other hand, 5 out of the 27 RPP samples in Table 1 are not listed in Table 2
194: for PWNe, because the following 5 pulsars have no PWN reported: B0628$-$28,
195: B0656+14, B0823+26, B0950+08 and B1055$-$52 ({\" O}gelman \& Tepedelenlio{\v g}lu
196: 2004, Becker et al. 2004, De Luca et al. 2005).
197:
198:
199:
200: \section{Analyses and Results}
201: The pulsar photon indices are distributed in a range of $1\la\Gamma_{\rm psr}\la3$ (as
202: shown in Fig \ref{fig:Lxvs.Gamma}). It should be noted that a significant fraction, about
203: $\sim15\%$ (4 out of 27), of the sources have soft spectra of $\Gamma_{\rm psr}>2$, which
204: may raise problems for current models as discussed later in \S 4. The photon indices of
205: the PWNe span a narrower range (Fig \ref{fig:Lxvs.Gamma}). As discussed below (\S 4),
206: this is consistent with the pulsar wind model.
207:
208: We investigate below the correlations between the X-ray emission properties of RPPs
209: and PWNe, and between the emission properties and the pulsar rotational parameters.
210: The rotation parameters include the period $P$, the period derivatives $\dot{P}$, and
211: some derived parameters, e.g., the magnetic field
212: $B=3.3\times10^{19}(P\dot{P})^{1/2}$G, the characteristic age $\tau=P/2\dot{P}$, and
213: the spin down power $\dot{E}=4\pi^2I\dot{P}/P^3$, where a typical moment of
214: $I=10^{45}\rm g\;cm^{2}$ is assumed. We have taken the values of $P$ and $\dot{P}$
215: from the pulsar catalog by Manchester et al. (2005)\footnote{See
216: http://www.atnf.csiro.au/research/pulsar/psrcat/}.
217:
218: In order to evaluate the significance level of the correlations of the two parameters
219: concerned, we calculate the widely used Pearson correlation coefficient ($r$), the
220: Spearman rank correlation coefficient ($r_{\rm s}$), and the two-sided significance level
221: ($p_{\rm s}$) of the Spearman rank test. The results are listed in Table 3.
222:
223: In addition to the correlation tests, we also perform a linear fit using the least
224: square method (LSM) to the relevant relations of the parameter pairs. Since the fitting
225: results are usually dominated by a few data points with much smaller observational errors
226: than the others, we also perform a linear fit without the observational errors for comparison.
227: The fitting results are all listed in Table 3, and shown in the relevant figures. In the
228: following we present the results in details.
229:
230: \subsection{Correlations between the RPP emission properties and the pulsar
231: rotational parameters}
232:
233: We study the RPP emission first. Strong correlations appear between the X-ray
234: luminosities of pulsars ($L_{\rm X, psr}$) and the pulsar rotational parameters (see
235: Table 3 and Figs \ref{fig_le} and \ref{fig_Lx_psr}).
236: First, $L_{\rm X,psr}$ is negatively correlated with $\tau$ and positively correlated
237: with $\dot{E}$, which are supported by the Spearman tests: $r_{\rm s} = -0.81$ and
238: $p_{\rm s}<0.0001$ between $L_{\rm X,psr}$ and $\tau$; and $r_{\rm s}=0.82$ and
239: $p_{\rm s}<0.0001$ between $L_{\rm X,psr}$ and $\dot{E}$.
240: We also note that there are some hints of the correlation hold for $L_{\rm X,psr}$ vs
241: $P$ and $\dot{P}$ separately, with the relevant Spearman rank correlation coefficients of
242: $-0.66$ and $0.69$ respectively and both significance levels $<0.001$.
243: The correlations between $L_{\rm X,psr}$ vs $P$ and $\dot{P}$ will disappear when the
244: sample includes both the MSPs and the normal RPPs, just as shown by Possenti et al.
245: (2002).
246:
247: Despite the Pearson and Spearman correlation coefficient support the existence of a
248: correlation between $L_{\rm X,psr}$ and $\dot{E}$ (or $\tau$), a simple linear fit to the
249: logarithm of the data points with the observational errors included does not produce a
250: statistically acceptable model. In fact, it results (here $L_{\rm X,psr}$ and $\dot{E}$
251: are in units of erg s$^{-1}$ and $\tau$ in years; see also Figs 2 and 3).
252: \[
253: L_{\rm X,psr}=10^{-0.8\pm1.3}\dot{E}^{0.92\pm0.04}(\chi^2 = 2.6) \]
254: \[ L_{\rm X,psr} = 10^{38.1\pm0.3}\tau^{-1.19\pm0.05}(\chi^2 = 4.9)
255: \]
256: Here and elsewhere in this paper, the uncertainties on the linear fits are reported at
257: 68\% confidence level. Previous authors (notably Possenti et al. 2002) also noticed that
258: a large scatter in the plot prevents to obtain an acceptable fit of the data with a
259: simple power law dependence of $L_{\rm X,psr}$ on $\dot{E}$. Hence this relation must
260: only be seen as an empirical average trend and not suitable for predicting the luminosity
261: of any specific source.
262:
263: We have also explored a linear fit which does not account for the uncertainties on the values of $L_{\rm X,psr}$. It turns out
264: \[
265: L^{*}_{\rm X,psr}=10^{-4.2\pm3.7}\dot{E}^{1.0\pm0.1} \]
266: \[ L^{*}_{\rm X,psr} = 10^{38.9\pm0.9}\tau^{-1.4\pm0.2}
267: \]
268: (see Figs \ref{fig_le} and~\ref{fig_Lx_psr}). A comparison between the current $L_{\rm
269: X,psr}$ versus $\dot{E}$ relation and the previous studies is shown in Fig \ref{fig_le}.
270: It can be seen that the relation we obtain above is close to the one between the pulsed
271: X-ray emission and the spin down power in Cheng et al. (2004), which indicates that most
272: of the non-thermal X-ray emission from a pulsar is pulsed.
273:
274: As already done by Possenti et al. (2002) using a sample including also the MSPs (but not
275: disentangling PWN from RPP emission), we also try to fit $L_{\rm X}$ with
276: $aP^{b}\dot{P}^c$. This gives the relation
277: \[
278: L_{\rm X,psr}=(40\pm1)P^{-3.4\pm0.3}\dot{P}^{0.77\pm0.07}(\chi^2=2.5).
279: \]
280: The nominal result of this (still statistically unacceptable) fit would suggest a
281: preferred dependence of $L_{\rm X,psr}$ on $\dot{E}/P$: however we note that, accounting
282: for the uncertainties on the parameters, the simpler dependence on $\dot{E}$ (recovered
283: in the work of Possenti et al. 2002) is also viable.
284:
285:
286: We also study the pulsar spectral properties, and check if there is any correlation
287: between the pulsar spectral index $\Gamma_{\rm psr}$ and the pulsar rotational
288: parameters. Inspection of Fig. \ref{fig_Gamma_psr}, may indicate the occurrence of a
289: positive correlation of $\Gamma_{\rm psr}$ with $P$ and $\tau$ and of a negative
290: correlation of $\Gamma_{\rm psr}$ with $\dot{P}$ and $\dot{E}$. However, a numerical
291: test indicates that all these correlations are too weak (the Spearman coefficients
292: $|r_{\rm s}|$ are all $\lesssim0.60$, see Table 3) for drawing any firm conclusion with
293: the available data.
294:
295: \subsection{Correlations between the PWN emission properties
296: and the pulsar rotational parameters}
297: We study here the correlations between the PWN X-ray
298: luminosity $L_{\rm X,pwn}$ and the pulsar rotational parameters. As shown in Table 3 and
299: Fig.~\ref{fig_Lx_pwn}, a weak positive correlation between $L_{\rm X,pwn}$ and $\dot{P}$ has
300: been detected.
301: Table 3 and Figs~\ref{fig_le} and~\ref{fig_Lx_pwn} also show that $L_{\rm X,pwn}$ is
302: strongly correlated with $\dot{E}$ and $\tau$.
303: The linear fits with the observational errors result
304: \[
305: L_{\rm X,pwn} = 10^{-19.6\pm3.0}\dot{E}^{1.45\pm0.08} (\chi^2=2.7)
306: \]
307: \[
308: L_{\rm X,pwn} = 10^{42.4\pm0.5}\tau^{-2.1\pm0.1} (\chi^2=5.0)
309: \]
310: Again, $L_{\rm X,pwn}$ and $\dot{E}$ are in units of erg~s$^{-1}$. Like for the emission
311: from RPPs, the adopted linear model in the logarithm of the data does not provide a
312: statistically acceptable fit.
313:
314: Trying a linear fit which does not account for the uncertainties on the values of $L_{\rm X,pwn}$, we obtain
315: \[
316: L^{*}_{\rm X,pwn} = 10^{-14.9\pm6.0}\dot{E}^{1.3\pm0.2}
317: \]
318: \[
319: L^{*}_{\rm X,pwn} = 10^{40.5\pm1.1}\tau^{-1.7\pm0.3}
320: \]
321: We note that the slope of the relation $L^{*}_{\rm X,pwn}\propto \dot{E}^{1.3}$ is
322: somewhat different from that of the pulsar, $L^{*}_{\rm X,psr}\propto\dot{E}$. The same
323: holds true comparing $L_{\rm X,pwn}\propto \dot{E}^{1.45}$ with $L_{\rm
324: X,psr}\propto\dot{E}^{0.92}$. It is worth noting that, as seen in Fig. \ref{fig_le}, the
325: scatterings in the relation of $L_{\rm X,psr}$ versus $\dot{E}$ and that of the PWNe are
326: comparable and both are large.
327:
328: As seen in Fig~\ref{fig_Gamma_pwn} and Table 3, there is no evidence for strong
329: correlations between $\Gamma_{\rm pwn}$ and the pulsar rotational parameters. The
330: Spearman rank test also supports this eye-ball study. However in Figs
331: \ref{fig:Lxvs.Gamma} and \ref{fig_Gpwn_lx_edot} we see obvious positive correlations
332: between the photon index $\Gamma_{\rm pwn}$ and the X-ray luminosity $L_{\rm X,pwn}$
333: or the X-ray conversion efficiency $L_{\rm X,pwn}/\dot{E}$. We will discuss the physical
334: implication in \S 4.
335:
336: \subsection{Correlations between non-thermal emission properties of RPPs and PWNe}
337: For those samples with both the RPP and the PWN non-thermal X-ray emission measured, we
338: test the correlations between them. As shown in Fig \ref{fig:pulvsPWN}, a strong
339: correlation between the X-ray luminosity of RPPs and that of PWNe appears in our samples,
340: while no correlation shown between their photon indices.
341:
342: The correlation test between two luminosities gives $r_{\rm s}=0.91$ and $p_{\rm
343: s}<10^{-4}$, and trying a linear fit to the relation leads to $L_{\rm
344: X,pwn}=10^{-1.9\pm3.2}L_{\rm X,psr}^{1.1\pm0.1}$. Given the strong positive
345: correlations of $L_{\rm X,psr}$ and $L_{\rm X,pwn}$ versus $\dot{E}$ separately, a
346: strong correlation between the two luminosities might be naturally expected. However,
347: the slope of the relation between the two luminosities is somewhat different from that
348: expected from the previous two relations, $L_{\rm X,psr}\propto\dot{E}$ and $L_{\rm
349: X,pwn}\propto\dot{E}^{1.3}$. This may be because that the samples used are different.
350: For example, those data points with $\dot{E}\la10^{34}\rm ergs\,s^{-1}$ are not
351: included for the relation of $L_{\rm X,pwn}$ versus $\dot{E}$ since no obvious PWNe
352: were detected, and they seem to result in a smaller slope for $L_{\rm X,psr}$ versus
353: $\dot{E}$ relation (Fig \ref{fig_le}). However, it should be noted that both $L_{\rm
354: X, psr}$ and $L_{\rm X, pwn}$ are actually modulated by the source distance.
355: The above correlation might be due to the effect of distance modulation.
356:
357: In our samples, the X-ray luminosity ratio between PWNe and RPPs, as shown by
358: $f_{\rm X,pwn}/f_{\rm X,psr}$ in Fig. \ref{fig:flux_ratio}, varies in the range of
359: $0.1-30$, about 2 orders of magnitude, and PWNe are generally brighter than their related
360: RPPs, typically $L_{\rm X,pwn}/L_{\rm X,psr}\sim1-10$. Fig. 9 also tells us that
361: a more energetic pulsar does not tend to transfer a bigger fraction
362: of $\dot{E}$ into PWN emission than to the pulsar emission, and vice versa.
363:
364: Gotthelf (2003) reported a linear relation between $\Gamma_{\rm psr}$ and $\Gamma_{\rm
365: pwn}$ for the Crab-like pulsars, but we find no correlation between them in our samples
366: (Fig. \ref{fig:pulvsPWN}). We also check the relation using the samples of Gotthelf
367: (2003), and a similar relation appears. These might suggest that the linear relation
368: exists only in the Crab-like pulsars, which are very young.
369:
370: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
371: \section{Discussions}
372:
373: \subsection{Non-thermal X-ray luminosities of RPPs and PWNe}
374: The detected X-ray emission from the RPPs and their PWNe is powered by the pulsar
375: rotation energy. In our sample, the conversion efficiency of $\dot{E}$ to the 2-10 keV
376: X-ray emission varies in the range of $(L_{\rm X,psr}+L_{\rm
377: X,pwn})/\dot{E}\sim10^{-6}-10^{-1}$, with the mean at $\sim10^{-3}$, so usually only a
378: small fraction of the spin-down power goes into the non-thermal X-ray emission. The
379: non-thermal X-ray emission in a source is usually dominated by the PWN rather than the
380: pulsar, and on average, $\langle L_{\rm X,pwn}/L_{\rm X,psr}\rangle\sim10$. This implies
381: that the relations of the pulsar X-ray luminosity and the spin-down power obtained in the
382: previous works using the low spatial resolution observations at $>2$keV might be
383: dominated by the PWN emission.
384:
385: In this work we can separately analyze the luminosity and the spin-down power
386: relations for the pulsars and their PWNe, thanks to the high spatial resolution of the
387: observations. A strong positive correlation between
388: $L_{\rm X,psr}$ and $\dot{E}$ is obtained in this paper,
389: similar to the previous results, as shown in Fig \ref{fig_le}. We compare our
390: results with the previous work in the following. However, one should keep in mind that
391: there are obvious differences in the analysis processes, i.e., we can separate the RPP
392: and the PWN emission and do not include the MSPs in the sample, while the previous
393: work did not separate the RPPs and the PWNe and included MSPs in the analysis.
394:
395: The relations for the X-ray luminosity of the RPPs in 2-10 keV band are $L_{\rm
396: X,psr}\propto \dot{E}^{0.92\pm0.04}$ (uncertainties on $L_{\rm X,psr}$ included) and
397: $L^{*}_{\rm X,psr}\propto \dot{E}^{1.0\pm0.1}$ (not accounting for the uncertainties on
398: $L_{\rm X,psr}$), respectively. They are roughly in agreement with the scaling found by
399: Becker \& Tr\"umper (1997), who used the X-ray luminosity in the 0.1-2.4 keV band. The
400: $L_{\rm X}-\dot{E}$ relations obtained by Saito(1998) and Possenti et al. (2002) appear
401: steeper than both our derived relations for RPPs, but they are roughly consistent with
402: both $L_{\rm X,pwn}\propto \dot{E}^{1.45\pm0.08}$ and $L^{*}_{\rm X,pwn}\propto
403: \dot{E}^{1.3\pm0.2}$ (Fig \ref{fig_le}). This may suggest that their relations could
404: also be influenced by the PWN emission due to the lower spatial resolution. The relations
405: $L_{\rm X,pul} \propto\dot{E}^{1.2\pm0.08}$ and $L_{\rm X,npul} \propto
406: \dot{E}^{1.4\pm0.1}$, obtained by Cheng et al. (2004) with {\sl ASCA} data, are similar
407: to our results.
408:
409: We have shown that the weak negative correlation of $L_{\rm X,psr}$ versus $P$ and the weak
410: positive correlation of $L_{\rm X,psr}$ versus $\dot{P}$ lead to a strong positive
411: correlation between $L_{\rm X,psr}$ and $\dot{E}$ in our MSP-excluding RPP samples. On
412: the other hand in the previous work including the MSPs and the normal RPPs, the
413: correlations between the X-ray luminosity (might include the PWN emission) and $P$ or
414: $\dot{P}$ disappear whereas a trend of $L_{\rm X}$ versus $\dot{E}$ is still there
415: (e.g., Possenti et al 2002). Moreover, the MSP samples alone also obey a similar
416: correlation (Possenti et al. 2002). All these factors together strongly suggest that the
417: X-ray luminosities of RPPs, including the MSPs, are only dependent on their spin-down
418: powers.
419:
420: Although there is a strong correlation between $L_{\rm X,psr}$ and $\dot{E}$, the
421: scattering in this relation is large and the linear fit with the observational errors
422: included usually gives a statistically unacceptable result, as suggested by Possenti et
423: al. (2002). $L_{\rm X,psr}$ at given $\dot{E}$ may spread over 2-4 orders of magnitude,
424: as seen in Fig \ref{fig_le}. The uncertainty in the distance determination and the
425: momenta of inertia are not expected to lead to such a large span, so other intrinsic
426: factors may work, e.g., the viewing angle effect, etc. The scattering in the $L_{\rm
427: X,pwn}-\dot{E}$ relation is comparably large, which is somewhat strange, since the PWN
428: emission is less influenced by the viewing angles.
429:
430:
431: \subsection{Non-thermal X-ray spectra of the RPPs}
432:
433: There are mainly two scenarios to produce the non-thermal X-rays in the magnetospheres of
434: pulsars. The outer gap scenario (e.g., Cheng et al. 1998; Wang et al. 1998; Cheng \&
435: Zhang 1999) produces a downward synchrotron-curvature cascade, where the secondary
436: electrons/positions produce X-rays by synchrotron emission. Another scenario is the polar
437: gap scenario, e.g., Zhang \& Harding (2000) proposed the ``full polar cap cascade", where
438: the non-thermal X-rays are produced by resonant inverse Compton (IC) scattering off the
439: thermal X-ray photons. In both scenarios the $L_{\rm X, psr}\propto\dot{E}$ relation is
440: generally predicted, although the X-ray spectra are not easy to understand.
441:
442: We note that a significant fraction of pulsars with very soft spectral indices,
443: $\Gamma_{\rm psr}\sim2-3$, may pose questions on current models. In a cascade, the
444: monoenergetic primary electrons emit monoenergetic curvature photons, which subsequently
445: turn into still monoenergetic pairs in a soft photon bath. The fast energy loss of the
446: secondary pairs in the magnetic field produces synchrotron emission, which have an photon
447: spectrum with a power law index $\Gamma_1=1.5$. If the cascade continues the photons
448: produce next-generation pairs and then synchrotron photons with index
449: $\Gamma_2=1+\Gamma_1/2=1.75$; furthermore, $\Gamma_3=1+\Gamma_2/2=1.875$... So the
450: indices will never be bigger than 2, in contrast with the soft spectra. Actually this
451: discussion could also work if IC rather than synchrotron emission is involved since the
452: index of synchrotron and IC emission is equal for the same energy distribution of pairs.
453: As for the polar gap scenarios, the synchrotron emission at X-rays is weak because the
454: secondary pairs with small pitch angles produce synchrotron emission well above the
455: cyclotron frequency in the strong pulsar magnetic field, typically $\sim100$~keV. Zhang
456: \& Harding (2000) proposed that the X-ray emission is dominated by the low energy tail in
457: the resonant IC emission. However this tail may be hard with index $\Gamma<2$, as shown
458: in some Monte Carlo simulations (e.g., Fang \& Zhang 2006), although the cases might be
459: more complicated when more factors such as the viewing angle are taken into account.
460:
461: Wang \& Zhao (2004) reported the possible negative correlations between
462: $\Gamma_{\rm psr}$ and $\dot{\Omega}$ and between $\Gamma_{\rm psr}$ and ${\zeta}$,
463: where $\zeta$ is the generation order parameter characterizing a pulsar under the
464: scheme of cascade processes (Zhao et al. 1989; Lu et al. 1994; Wei et al. 1997). A
465: similar negative correlation between $\Gamma_{\rm psr}$ and ${\zeta}$ in anomalous
466: X-ray pulsars and softer gamma-ray repeaters had also been reported (Marsden \& White
467: 2001; Lu et al. 2003), suggesting that a common mechanism may operate in both normal
468: and anomalous pulsars. These observational results seem in contrast with the predicted
469: positive correlation between $\Gamma_{\rm psr}$ and $\zeta$ by Lu et al. (1994). Here,
470: we also check the relation between $\Gamma_{\rm psr}$ and the generation order
471: parameter $\zeta_3= 1 + (0.6 - (11/14){\rm log}P + (2/7){\rm log}\dot{P}_{15})/1.3$
472: (Eq. 6 of Wang \& Zhao 2004) and list the correlation results in Table 3. It turns out
473: that although there may be some hints of such a negative correlation in our sample,
474: the correlation tests do not support it strongly, $r_{\rm s}=-0.5$. So the current
475: data are not good enough to test the theoretical predictions of Lu et al. (1994).
476:
477: \subsection{X-ray spectra of PWNe}
478:
479: In the standard Kennel \& Coroniti (1984ab) model for the Crab nebula, the young Crab
480: pulsar loses its rotational energy predominantly in the form of a highly relativistic
481: particle wind, which encounters with the surrounding medium in a termination shock and
482: become visible by synchrotron emission downstream from the shock. In this context
483: the energy in the relativistic wind is transferred into post shock magnetic field and
484: accelerated particles with energy distribution of $N_e(E_e)\propto E_e^{-p}$. Chevalier
485: (2000) discussed the PWN spectra with emphasis on the cooling of the X-ray emitting
486: electrons, which leads to a steeper index $p+1$ for high energy and fast-cooling electrons,
487: and hence a spectral transition of synchrotron photons from $(p+1)/2$ in the slow-cooling
488: regime to $(p+2)/2$ in the fast-cooling regime at break frequency ($\nu_c$).
489:
490: The data fit with a single power law model to the indeed broken power law would result in a
491: spectral index always in the range of $(p+1)/2\leq\Gamma_{\rm pwn}\leq(p+2)/2$. An
492: electron index value $p\approx2.2$ is generally obtained in theoretical works on particle
493: acceleration in relativistic collisionless shocks, by both numerical calculations (e.g.,
494: Achterberg et al. 2001) and analytic analysis (e.g., Keshet \& Waxman 2005), and also
495: inferred from observation in other kinds of astrophysical relativistic shocks, e.g., GRB
496: afterglows (e.g., Freedman \& Waxman 2001). Our results show an narrow index range of
497: $1.5\la\Gamma_{\rm pwn}\la2.1$ unless one source with somewhat higher value $\sim2.5$,
498: suggesting an electron index of $p\sim2.2$. This consistence with particle wind models
499: gives a strong support to the Fermi-shock acceleration in PWNe.
500:
501: We show that the PWN spectral parameters are not strongly correlated with the pulsar
502: rotational parameters (Fig \ref{fig_Gamma_pwn}). Gotthelf (2003) reported the correlation
503: between $\Gamma_{\rm psr}$ and $\Gamma_{\rm pwn}$ for nine Crab-like pulsars. Our studies
504: show that such a correlation is probably not a common property for all RPPs. Therefore,
505: the electron spectrum and its evolution in a PWN are not determined by the central
506: pulsar, consistent with wind models where the emission comes from a relativistic shock
507: between wind and environment interaction.
508:
509: The relation of $\Gamma_{\rm pwn}$ with PWN luminosity $L_{\rm X,pwn}$ and the
510: conversion efficiency $L_{\rm X,pwn}/\dot{E}$ (Figs~\ref{fig:Lxvs.Gamma} and
511: \ref{fig_Gpwn_lx_edot} and Table 3) could be understood qualitatively in the framework of
512: pulsar wind models taking into account the electron cooling effect on spectral profile
513: (e.g., Chevalier 2000). If pulsar loses most of its rotation energy through particle
514: winds, then higher $\dot{E}$ corresponds to stronger cooling and hence lower spectral
515: break $\nu_c$, which further means a larger index $\Gamma_{\rm pwn}$ in a fixed
516: observational energy range. In the meantime, a higher $\dot{E}$ corresponds to a larger
517: $L_{\rm X,pwn}$, no matter $\nu_c$ is below or above the observational range, and
518: corresponds to constant X-ray conversion efficiency for fast cooling regime ($\nu_c$
519: below observed range) or larger $L_{\rm X,pwn}/\dot{E}$ for slow cooling regime ($\nu_c$
520: above observed range). Therefore we have softer PWN spectra (larger $\Gamma_{\rm pwn}$)
521: for more luminous PWNe (larger $L_{\rm X,pwn}$) and higher energy conversion efficiency
522: ($L_{\rm X,pwn}/\dot{E}$). This consistence supports the wind-shock model for PWNe. In
523: this context, the transition of $\Gamma_{\rm pwn}$ from high to low values in Fig.
524: \ref{fig:Lxvs.Gamma} suggests that the spectral break $\nu_c$ locates at 2-10 keV for
525: $L_{\rm X,pwn}\sim10^{33}\rm ergs\, s^{-1}$. This may give constraint to wind model
526: parameters.
527:
528:
529: \section{Conclusions}
530:
531: In this work, using the available samples of 27 RPPs and 24 PWNe observed by {\sl
532: Chandra} and {\sl XMM-Newton}, we obtain the non-thermal X-ray spectral properties, i.e.,
533: luminosities and spectral indices, of RPPs and PWNe separately. We then analyze their
534: distribution and correlation with each other and with pulsar rotational parameters.
535: \begin{itemize}
536:
537: \item As to the correlations we find: (1) $L_{\rm X,psr}$ and $L_{\rm X,pwn}$
538: display a strong correlation with both $\dot{E}$ and $\tau$; (2) $L_{\rm X,psr}$ also
539: shows a possible weaker correlation with $P$ and $\dot{P}$, whereas $L_{\rm X,pwn}$
540: manifests a similar weak correlation with $\dot{P}$ only; (3) $\Gamma_{\rm pwn}$ is
541: positively correlated with $L_{\rm X,pwn}$ and the efficiency of conversion of rotational
542: energy loss in X-ray luminosity $L_{\rm X,pwn}/\dot{E}$.
543:
544: \item Trying to fit the logarithm of the data with a simple linear fit, we find: $L_{\rm
545: X,psr}=10^{-0.8\pm1.3}\dot{E}^{0.92\pm0.04}$ and $ L_{\rm X,pwn} =
546: 10^{-19.6\pm3.0}\dot{E}^{1.45\pm0.08}$. However, both the fits are statistically
547: unacceptable. Not accounting for the uncertainties on the observed luminosity, the
548: aforementioned relations become $L^{*}_{\rm X,psr}=10^{-4.2\pm3.7}\dot{E}^{1.0\pm0.1} $ and
549: $L^{*}_{\rm X,pwn} = 10^{-14.9\pm6.0}\dot{E}^{1.3\pm0.2}$, respectively.
550: Since the scatter in the relation
551: for PWN (whose emission should be less affected by viewing angle) is comparably larger
552: than that for RPPs, the scatter in the relation is more probably intrinsic to the
553: sources.
554:
555:
556: \item The PWN X-ray luminosity is typically 1 to 10 times larger than that from the
557: underlying pulsar.
558:
559:
560: \item The pulsar photon index spans a range of $1\la\Gamma_{\rm
561: psr}\la3$. A significant fraction of RPPs with low $\dot{E}$ show soft spectra of $\Gamma_{\rm psr}>2$, which seems not
562: consistent with the current models and urges for further investigation of the non-thermal X-ray emission mechanisms of pulsars.
563:
564: \item The PWN spectral properties are consistent with the particle wind model: the photon
565: index range $1.5\la\Gamma_{\rm pwn}\la2$ is
566: consistent with that expected from the shock-accelerated electrons of index $p\sim2$;
567: the correlations of $\Gamma_{\rm pwn}$ with
568: $L_{\rm X,pwn}$ and the conversion efficiency $L_{\rm X,pwn}/\dot{E}$ are
569: consistent with the wind
570: model; no correlation between $\Gamma_{\rm pwn}$ and the pulsar rotational parameters
571: also implies that the cooling process is not related to the center pulsars but to the
572: interaction of the pulsar wind with its environment.
573:
574: \end{itemize}
575:
576: \section*{Acknowledgments}
577: We thank the referee for very thorough comments. We are grateful to S. N. Zhang, L. M.
578: Song, J. M. Wang, G. J. Qiao, B. Zhang and H. G. Wang for helpful discussion. We thank
579: Prof. Wielebinski and Dr. Jessner for critically reading the manuscript and giving many
580: valuable suggestions. XHL sincerely thanks Prof. Wielebinski for the financial support
581: during her stay at MPIfR, and thanks J.L. Han for warm hospitality during her stay at
582: NAOC. This work is supported by the National Science Foundation of China through grants
583: 10573017, 10533020 and 10473010.
584:
585: \begin{thebibliography}{99}
586: \bibitem[Achterberg et al. (2001)]{Achterberg2001}
587: Achterberg, A., Gallant, Y.~A., Kirk, J.~G., \& Guthmann, A.~W.\ 2001, \mnras,
588: 328, 393
589:
590: \bibitem[Becker \& Tr{\"u}mper(1997)]{1997A&A...326..682B} Becker, W., \&
591: Tr{\"u}mper, J.\ 1997, A\&A, 326, 682
592:
593: \bibitem[Becker et al.(2004)]{2004ApJ...615..908B} Becker, W., Weisskopf,
594: M.~C., Tennant, A.~F., Jessner, A., Dyks, J., Harding, A.~K., \& Zhang,
595: S.~N.\ 2004, ApJ, 615, 908
596:
597: \bibitem[Bradley (2007)]{2007arXiv0709.4531B} Bradley, E. Schaefer
598: \ 2007, ArXiv e-prints, 709, arXiv:0709.4531
599:
600: \bibitem[Brisken et al.(2002)]{2002ApJ...571..906B} Brisken, W.~F., Benson,
601: J.~M., Goss, W.~M., \& Thorsett, S.~E.\ 2002, \apj, 571, 906
602:
603: \bibitem[Brisken et al.(2003)]{2003ApJ...593L..89B} Brisken, W.~F.,
604: Thorsett, S.~E., Golden, A., \& Goss, W.~M.\ 2003, \apjl, 593, L89
605:
606: \bibitem[Camilo et al.(2002)]{Camilo2002} Camilo, F., Lorimer,
607: D.~R., Bhat, N.~D.~R., Gotthelf, E.~V., Halpern, J.~P., Wang, Q.~D., Lu,
608: F.~J., \& Mirabal, N.\ 2002, ApJL, 574, L71
609:
610: \bibitem[Camilo et al.(2004)]{2004ApJ...616.1118C} Camilo, F., Gaensler,
611: B.~M., Gotthelf, E.~V., Halpern, J.~P., \& Manchester, R.~N.\ 2004, \apj,
612: 616, 1118
613:
614: \bibitem[Caraveo et al.(2004)]{2004Sci...305..376C} Caraveo, P.~A., De
615: Luca, A., Mereghetti, S., Pellizzoni, A., \& Bignami, G.~F.\ 2004, Science,
616: 305, 376
617:
618: \bibitem[Caraveo et al.(2001)]{2001ApJ...561..930C} Caraveo, P.~A., De
619: Luca, A., Mignani, R.~P., \& Bignami, G.~F.\ 2001, \apj, 561, 930
620:
621: \bibitem[Chatterjee et al.(2004)]{Chatterjee2004} Chatterjee, S.,
622: Cordes, J.~M., Vlemmings, W.~H.~T., Arzoumanian, Z., Goss, W.~M., \& Lazio,
623: T.~J.~W.\ 2004, \apj, 604, 339
624:
625: \bibitem[Cheng et al.(1998)]{Cheng1998} Cheng, K.~S., Gil, J., \&
626: Zhang, L.\ 1998, \apjl, 493, L35
627:
628: \bibitem[Cheng et al.(2004)]{Cheng2004} Cheng, K.~S., Taam,
629: R.~E., \& Wang, W.\ 2004, ApJ, 617, 480
630:
631: \bibitem[Cheng \& Zhang(1999)]{Cheng1999} Cheng, K.~S., \& Zhang,
632: L.\ 1999, \apj, 515, 337
633:
634: \bibitem[Chevalier(2000)]{2000ApJ...539L..45C} Chevalier, R.~A.\ 2000,
635: \apjl, 539, L45
636:
637: \bibitem[Cordes \& Lazio(2002)]{cordes2002} Cordes, J. M., \& Lazio, T. J. W. 2002, preprint (astro-ph/0207156)
638:
639: \bibitem[Cox et al.(1999)]{1999ApJ...524..179C} Cox, D.~P., Shelton, R.~L.,
640: Maciejewski, W., Smith, R.~K., Plewa, T., Pawl, A., \& R{\'o}{\.z}yczka,
641: M.\ 1999, \apj, 524, 179
642:
643: \bibitem[De Luca et al.(2005)]{2005ApJ...623.1051D} De Luca, A., Caraveo,
644: P.~A., Mereghetti, S., Negroni, M., \& Bignami, G.~F.\ 2005, ApJ, 623,
645: 1051
646:
647: \bibitem[Faherty et al.(2007)]{2007Ap&SS.308..225F} Faherty, J., Walter,
648: F.~M., \& Anderson, J.\ 2007, \apss, 308, 225
649:
650: \bibitem[Fang \& Zhang(2006)]{2006ApJ...653..573F} Fang, J., \& Zhang, L.\
651: 2006, \apj, 653, 573
652:
653: \bibitem[Freedman \& Waxman (2001)]{Freedman01} Freedman, D.~L., \&
654: Waxman, E.\ 2001, \apj, 547, 922
655:
656: \bibitem[Gaensler \& Wallace(2003)]{2003ApJ...594..326G} Gaensler, B.~M.,
657: \& Wallace, B.~J.\ 2003, \apj, 594, 326
658:
659: \bibitem[Gaensler et al.(2004)]{2004ApJ...616..383G} Gaensler, B.~M., van
660: der Swaluw, E., Camilo, F., Kaspi, V.~M., Baganoff, F.~K., Yusef-Zadeh, F.,
661: \& Manchester, R.~N.\ 2004, \apj, 616, 383
662:
663: \bibitem[Gil \& Krawczyk(1996)]{1996MNRAS.280..143G} Gil, J., \& Krawczyk,
664: A.\ 1996, \mnras, 280, 143
665:
666: \bibitem[Gotthelf et al.(2002)]{Gotthelf2002} Gotthelf, E.~V.,
667: Halpern, J.~P., \& Dodson, R.\ 2002, ApJ, 567, L125
668:
669: \bibitem[Gotthelf(2003)]{2003ApJ...591..361G} Gotthelf, E.~V.\ 2003, ApJ,
670: 591, 361
671:
672: \bibitem[Green et al.(1988)]{1988MNRAS.231..735G} Green, D.~A., Gull,
673: S.~F., Tan, S.~M., \& Simon, A.~J.~B.\ 1988, \mnras, 231, 735
674:
675: \bibitem[Hessels et al.(2004)]{Hessels2004} Hessels, J.~W.~T.,
676: Roberts, M.~S.~E., Ransom, S.~M., Kaspi, V.~M., Romani, R.~W., Ng, C.-Y.,
677: Freire, P.~C.~C., \& Gaensler, B.~M.\ 2004, ApJ, 612, 389
678:
679: \bibitem[Hui \& Becker(2007)]{2007A&A...470..965H} Hui, C.~Y., \& Becker,
680: W.\ 2007, \aap, 470, 965
681:
682: \bibitem[Kargaltsev \& Pavlov(2007)]{2007ApJ...670..655K} Kargaltsev, O.,
683: \& Pavlov, G.~G.\ 2007, \apj, 670, 655
684:
685: \bibitem[Kaspi et al.(2001)]{Kaspi2001} Kaspi, V.~M., Gotthelf,
686: E.~V., Gaensler, B.~M., \& Lyutikov, M.\ 2001, ApJL, 562, L163
687:
688: \bibitem[Kennel \& Coroniti(1984)]{1984ApJ...283..694K} Kennel, C.~F., \&
689: Coroniti, F.~V.\ 1984a, \apj, 283, 694
690:
691: \bibitem[Kennel \& Coroniti(1984)]{1984ApJ...283..710K} Kennel, C.~F., \&
692: Coroniti, F.~V.\ 1984b, \apj, 283, 710
693:
694: \bibitem[Keshet \& Waxman (2005)]{2005PhRvL..94k1102K} Keshet, U., \&
695: Waxman, E.\ 2005, Physical Review Letters, 94, 111102
696:
697: \bibitem[Koribalski et al.(1995)]{1995ApJ...441..756K} Koribalski, B.,
698: Johnston, S., Weisberg, J.~M., \& Wilson, W.\ 1995, \apj, 441, 756
699:
700: \bibitem[Kothes et al.(2006)]{Kothes2006} Kothes, R., Reich, W.,
701: \& Uyan{\i}ker, B.\ 2006, \apj, 638, 225
702:
703: \bibitem[Kramer et al.(2003)]{2003MNRAS.342.1299K} Kramer, M., et al.\
704: 2003, \mnras, 342, 1299
705:
706: \bibitem[Kundt \& Schaaf(1993)]{1993Ap&SS.200..251K} Kundt, W., \& Schaaf,
707: R.\ 1993, \apss, 200, 251
708:
709: \bibitem[Leahy
710: \& Tian(2007)]{2007arXiv0711.4107L} Leahy, D.~A., \& Tian, W.\ 2007, ArXiv e-prints, 711, arXiv:0711.4107
711:
712: \bibitem[Lu et al.(2002)]{Lu2002} Lu, F.~J., Wang, Q.~D.,
713: Aschenbach, B., Durouchoux, P., \& Song, L.~M.\ 2002, ApJL, 568, L49
714:
715: \bibitem[Lu et al.(1994)]{Lu1994} Lu, T., Wei, D.M., \& Song, L.M.\ 1994, A\&A, 290, 815
716:
717: \bibitem[Lu et al.(2003)]{2003ChJAA...3..543L} Lu, Y., Wang, W., \& Zhao,
718: Y.-H.\ 2003, Chinese Journal of Astronony and Astrophysics, 3, 543
719:
720: \bibitem[Manchester et al.(2005)]{2005AJ....129.1993M} Manchester, R.~N.,
721: Hobbs, G.~B., Teoh, A., \& Hobbs, M.\ 2005, \aj, 129, 1993
722:
723: \bibitem[Marsden \& White(2001)]{Marsden2001} Marsden, D., \&
724: White, N.~E.\ 2001, \apjl, 551, L155
725:
726: \bibitem[{\" O}gelman \& Tepedelenlio{\v g}lu(2004)]{2004AdSpR..33..597O}
727: {\" O}gelman, H., \& Tepedelenlio{\v g}lu, E.\ 2004, Advances in Space
728: Research, 33, 597
729:
730: \bibitem[Paron et al.(2006)]{2006PASA...23...69P} Paron, S.~A., Reynoso,
731: E.~M., Purcell, C., Dubner, G.~M., \& Green, A.\ 2006, Publications of the
732: Astronomical Society of Australia, 23, 69
733:
734: \bibitem[Pavlov et al.(2001)]{2001ApJ...552L.129P} Pavlov, G.~G., Zavlin,
735: V.~E., Sanwal, D., Burwitz, V., \& Garmire, G.~P.\ 2001, \apjl, 552, L129
736:
737: \bibitem[Petre et al.(2002)]{Petre2002} Petre, R., Kuntz, K.~D.,
738: \& Shelton, R.~L.\ 2002, ApJ, 579, 404
739:
740: \bibitem[Possenti et al.(2002)]{2002A&A...387..993P} Possenti, A., Cerutti,
741: R., Colpi, M., \& Mereghetti, S.\ 2002, A\&A, 387, 993
742:
743: \bibitem[Roberts et al.(1993)]{1993A&A...274..427R} Roberts, D.~A., Goss,
744: W.~M., Kalberla, P.~M.~W., Herbstmeier, U., \& Schwarz, U.~J.\ 1993, \aap,
745: 274, 427
746:
747: \bibitem[Roberts et al.(2003)]{Roberts2003} Roberts, M.~S.~E., Tam,
748: C.~R., Kaspi, V.~M., Lyutikov, M., Vasisht, G., Pivovaroff, M., Gotthelf,
749: E.~V., \& Kawai, N.\ 2003, ApJ, 588, 992
750:
751: \bibitem[Ruiz \& May(1986)]{1986ApJ...309..667R} Ruiz, M.~T., \& May, J.\
752: 1986, \apj, 309, 667
753:
754: \bibitem[Saito (1998)]{Saito1998} Saito, Y. 1998, Ph.D. Thesis, Univ. of Tokyo (S98)
755:
756: \bibitem[Seward \& Wang(1988)]{1988ApJ...332..199S} Seward, F.~D., \& Wang,
757: Z.-R.\ 1988, ApJ, 332, 199
758:
759: \bibitem[Strom \& Stappers(2000)]{2000ASPC..202..509S} Strom, R.~G., \&
760: Stappers, B.~W.\ 2000, IAU Colloq.~177: Pulsar Astronomy - 2000 and Beyond,
761: 202, 509
762:
763: \bibitem[Tennant et al.(2001)]{Tennant2001} Tennant, A.~F., et al.\
764: 2001, \apjl, 554, L173
765:
766: \bibitem[Trimble \& Woltjer(1971)]{1971ApJ...163L..97T} Trimble, V., \&
767: Woltjer, L.\ 1971, \apjl, 163, L97
768:
769: \bibitem[Wang et al.(1998)]{Wang1998} Wang, F.~Y.-H., Ruderman,
770: M., Halpern, J.~P., \& Zhu, T.\ 1998, \apj, 498, 373
771:
772: \bibitem[Wang \& Zhao(2004)]{Wang2004} Wang, W., \& Zhao, Y.\
773: 2004, \apj, 601, 1038
774:
775: \bibitem[Wei et al.(1997)]{Wei1997} Wei, D.M., Song, L.M., \& Lu, T.\ 1997,
776: A\&A, 323, 98
777:
778: \bibitem[Willingale et al.(2001)]{2001A&A...365L.212W} Willingale, R.,
779: Aschenbach, B., Griffiths, R.~G., Sembay, S., Warwick, R.~S., Becker, W.,
780: Abbey, A.~F., \& Bonnet-Bidaud, J.-M.\ 2001, \aap, 365, L212
781:
782: \bibitem[Zavlin et al.(1995)]{1995AstL...21..149Z} Zavlin, V.~E., Shibanov,
783: Y.~A., \& Pavlov, G.~G.\ 1995, Astronomy Letters, 21, 149
784:
785: \bibitem[Zhang \& Harding(2000)]{2000ApJ...532.1150Z} Zhang, B., \&
786: Harding, A.~K.\ 2000, \apj, 532, 1150
787:
788: \bibitem[Zhao et al.(1989)]{Zhao1989}Zhao, Y.H., Huang, K.L., Peng, Q.H.,
789: Lu, T., \& Lu, J.L.\ 1989, A\&A, 223, 147
790:
791: \end{thebibliography}
792: \bibliographystyle{aa}
793:
794:
795: \include{table1}
796: \include{table2}
797: \include{table3}
798:
799: \begin{figure}[!htbp]
800: \includegraphics[scale=.6]{f1a.ps}
801: \includegraphics[scale=.6]{f1b.ps}
802: \caption{ The relations between the non-thermal X-ray luminosity
803: in 2-10 keV and the photon index for RPPs (left panel) and PWNe (right panel). The
804: solid lines are the best LSM fit without observational errors taken into account, while the
805: dotted line with observational errors.
806: The dashed lines mark the cases of $\Gamma$=2 for comparisons.} \label{fig:Lxvs.Gamma}
807: \end{figure}
808:
809:
810: \begin{figure}
811: \includegraphics[scale=.6]{f2a.ps}
812: \includegraphics[scale=.6]{f2b.ps}
813: \caption{ Left: The
814: non-thermal X-ray luminosities ($L_{\rm X,psr}$) in 2-10 keV from the {\sl Chandra} and {\sl
815: XMM-Newton} observations vs. spin-down power $\dot{E}$ of 27 pulsars. Right:
816: The non-thermal X-ray luminosity ($L_{\rm X,pwn}$) in 2-10 keV from the {\sl Chandra} and
817: {\sl XMM-Newton} observations vs. spin-down power $\dot{E}$ of 24 PWNe. The bold
818: solid line is the best fit to the data by LSM without observational errors included, while
819: the thin solid line is the fit with observational errors. For
820: comparison the resulted relations in previous works are also marked: the dash-dot-dot-dot
821: line corresponds to results of Becker \& Tr{\"u}mper (1997); the dash-dot line to Saito
822: (1998); the short dash line to Possenti et al. (2002); and long dash line to Cheng et al.
823: (2004).} \label{fig_le}
824: \end{figure}
825:
826:
827: \begin{figure}
828: \includegraphics[scale=1.]{f3.ps}
829: \caption{Relations of $L_{\rm X,psr}$ vs. $P$, $\dot{P}$ and $\tau$. The solid lines are
830: the best LSM fit without observational errors considered while the dotted lines are with
831: errors. The fitting results are also listed in Table 3.} \label{fig_Lx_psr}
832: \end{figure}
833:
834: \begin{figure}
835: \includegraphics[scale=.7]{f4.ps}
836: \caption{Relations of $\Gamma_{\rm psr}$ vs. $P$, $\dot{P}$, $\tau$ and $\dot{E}$. The
837: solid lines are the best LSM fit without observational errors considered while the dotted
838: lines are with errors. The fitting results are also listed in Table 3.}
839: \label{fig_Gamma_psr}
840: \end{figure}
841:
842:
843:
844: \begin{figure}
845: \includegraphics[scale=1.]{f5.ps}
846: \caption{Relations of $L_{\rm X, pwn}$ vs. $P$, $\dot{P}$ and $\tau$. The solid lines are
847: the best LSM fit without observational errors considered while the dotted lines are with
848: errors. The fitting results are also listed in Table 3.} \label{fig_Lx_pwn}
849: \end{figure}
850:
851: \begin{figure}
852: \includegraphics[scale=.7]{f6.ps}
853: \caption{Relations of $\Gamma_{\rm pwn}$ vs. $P$, $\dot{P}$, $\tau$ and $\dot{E}$. The
854: solid lines are the best LSM fit with observational errors considered while the dotted
855: lines are without errors. The fitting results are also listed in Table 3.}
856: \label{fig_Gamma_pwn}
857: \end{figure}
858:
859: \begin{figure}
860: \includegraphics[scale=.6]{f7.ps}
861: \caption{The relation between $\Gamma_{\rm pwn}$ and $L_{\rm X,pwn}/\dot{E}$. The solid
862: line is the best LSM fit without observational errors considered while the dotted line is
863: with errors. The fitting results are also listed in Table 3.} \label{fig_Gpwn_lx_edot}
864: \end{figure}
865:
866: \begin{figure}
867: \includegraphics[scale=.6]{f8a.ps}
868: \includegraphics[scale=.6]{f8b.ps}
869: \caption{Left: the relation of $L_{\rm X, psr}$ vs. $L_{\rm X, pwn}$. The solid lines are
870: the best LSM fit with the observational errors considered while the dotted lines are
871: without errors. The fitting results are also listed in Table 3. The dash line marking the
872: case of $L_{\rm X, psr}=L_{\rm X, pwn}$ is for comparison. Right: the relation of
873: $\Gamma_{\rm psr}$ vs. $\Gamma_{\rm pwn}$. } \label{fig:pulvsPWN}
874: \end{figure}
875:
876: \begin{figure}
877: \includegraphics[scale=.6]{f9.ps}
878: \caption{ The relation of $f_{\rm X, pwn}/f_{\rm X, psr}$ vs $\dot{E}$. The dash line
879: indicates $f_{\rm X, pwn} = f_{\rm X, psr}$.} \label{fig:flux_ratio}
880: \end{figure}
881:
882:
883: \end{document}
884: