arXiv:astro-ph/0305383v1 20 May 2003 2 3 h rgntrsa eg,Hge ig 94.Recent 1994). with Singh A of Cas & mass SNR the Hughes the for of (e.g., studies range a star estimate the progenitor SN, in su- the collapse and, the core thermonuclear) a of of or nature case the collapse constrain core to (e.g., used pernova be can (SNR) nant 1 n vlto fplasadterwn eua (PWNe). origin nebulae the wind to their and prep.) in of 2003, evolution from derived and opportu- al. star et the progenitor (Park the have on nucleosynthesis we information tie G292.0+1.8 to With nity ( nearby 2003). a relatively Wallace having being 2002); and al. et 1979); ( Park age determined X-ray: dynamically 1979; (optical: Clark ejecta synthe- & magnesium-rich newly Murdin and for by neon-, evidence associations oxygen-, spectacular sized /SNR showing other of all virtue G292.0+1.8 nearly W44). from and Vela differs from (e.g., de- distinguished gas for easily interstellar be distant shocked cannot too ejecta shocked the are that for 58), E0540 evolved SNR 3C evidence (e.g., and much study Crab tailed show inves- the not such (e.g., for do ejecta useless they virtually that are tigations: SNRs obser- pulsars most Willingale new Unfortunately, young the 2001; harbor studies. of al. potential such great et for the vatories Bleeker highlight 2002) 2000; al. al. et et Hughes (e.g., on usrwti rna 22018uigteParkes the using 135-ms G292.0+1.8 a near discovered or a (2002) within In al. pulsar 2001). young et al. Camilo et study (Hughes follow-up G292.0+1.8 in nebula chrotron J112439.1 (CXOU rpittpstuigL using typeset 2003 Preprint May 20 of version Draft avr-mtsna etrfrAtohsc,6 adnSt Garden 60 , for Center Harvard-Smithsonian eateto srnm n srpyis enyvnaSt Pennsylvania Astrophysics, and Astronomy of Department eateto hsc n srnm,RtesUiest,1 University, Rutgers Astronomy, and of Department h opsto fteeet eni uenv rem- a in seen ejecta the of composition The Recently oln uvs hr sn pia onepr otenwpulsar; new pulsar. the Crab to the headings: b of counterpart Subject slightly that optical or below no at, 5 ene is are of minimum star factor There the neutron from curves. the far of cooling is issu temperature that lifetime surface as PWN well the particle-dominated as a nebula, for synchrotron larger argues the and wind pulsar s -a Observatory X-ray ekdadbodwt WMwdho 0.23 of width FWHM a with broad and peaked ihntesproarmat(N)G9.+. sn h ihResolu High the using G292.0+1.8 (SNR) remnant supernova the within 3 us eido S J1124 PSR of period pulse h usraewl ecie yafaueespwrlwmdlwt a with model law power featureless a by described well are pulsar the − . 1 erpr h icvr fple -a msinfo h compact the from emission X-ray pulsed of discovery the report We 1 × epasbyietf h oaino h usrstriainsok P shock. termination pulsar’s the of location the identify plausibly We . Chandra onP Hughes P. John 10 NXRYPLA NTEOYE-IHSPROARMATG292.0+ REMNANT SUPERNOVA OXYGEN-RICH THE IN PULSAR X-RAY AN 21 cm − A 1. T 960 etrdo iuesyn- diffuse a on centered 591620) E − tl emulateapj style X introduction eelda -a on source point X-ray an revealed 2 htnidxΓ=1 = Γ index photon , PRJ1124 (PSR CO J112439.1 (CXOU h -a eid( period X-ray The . S:idvda SRG9.+.,MH11 MSH G292.0+1.8, (SNR individual ISM: − ∼ 1 Chandra arc .Slane O. Patrick , 93adN5B,o r so are or N157B), and 69.3 00ys udn&Clark & Murdin yrs; 2000 − ∼ 96frasidw aeof rate spindown a for 5916 − p;Gese & Gaensler kpc; 6 and 96 tr:nurn–sproarmat -as individua X-rays: – remnants supernova – neutron stars: – 5916) XMM-Newton Received − . ,aduasre .–0kVbn luminosity band keV 0.3–10 unabsorbed and 6, 591620) P 2 0 = agokPark Sangwook , et abig,M 23;[email protected] 02138; MA Cambridge, reet, .Burrows N. 6FeigusnRa,Psaaa,N 85-09 jph@phy 08854-8019; NJ Piscataway, Road, Frelinghuysen 36 ABSTRACT t nvriy 2 ae aoaoy nvriyPr,PA. Park, University Laboratory, Davey 525 University, ate . 3395s scnitn ihetaoaino h radio the of extrapolation with consistent is s) 13530915 P (83 1 accepted ; ◦ .Tepleaeae -a pcrlpoete of properties spectral X-ray pulse-averaged The ). .188 ,00530 ,ad00265 corresponding s 0.00296954 of and s, s, duration 0.0237563 0.00593907 of entire s, events. binsizes 0.0118781 the different these four for on using observation constructed pulsations the were for search curves we blind First Light a photons. X-ray out 1324 carried detected we source point the hpdnbl.Wti aiso 2 of radius a elliptically- Within diffuse, nebula. small, a shaped on centered source = unresolved decl. position 11:24:39.1, at = candidate pulsar R.A. the containing image HRC the edo iwruhy6 roughly provides view which of array plane field cen- focal a the HRC-S the utilize aimpoint of observations the portion mode tral at Timing candidate pulsar 1953). the (ObsID with mode timing in ai eecp hti lotsrl h onepr to counterpart the surely almost is the that telescope radio akstlsoe( telescope Parkes h ieiecretdepsr a 97 s. 49578 was of exposure gaps corrected livetime 8 The were interruptions was observation only ∼ Our the candidate. continuous; pulsar using nearly the barycenter of solar-system 16 position the about the to of times accuracy arrival an photon to tagged time 01uigthe using 2001 iyn ta h opc enn fteS htformed that SN the of remnant compact the G292.0+1.8. as the spa- it detect and tifying we J112439.1 temporal CXOU which from high in signal on observations pulsed X-ray report resolution we tial Here ironclad. not uainec,dsrbtdtruhu h exposure. the throughout distributed each, duration s 2 iue1(etpnl hw ogl 6 roughly a shows panel) (left 1 Figure eosre N 22018bgnigo 4July 14 on beginning G292.0+1.8 SNR observed We P 3 ˙ Chandra 7 = 3 ee .A Roming A. W. Peter , . − 6 lw ausepce rm“standard” from expected values elow, 5 × 4 usr:individual pulsars: – ) 10 on ore oee h ag emo the of beam large the However source. point sfrteXryeitn electrons, X-ray-emitting the for es t pia uioiyi tlata least at is luminosity optical its bobn oundensity column absorbing n − Chandra betCO J112439.1 CXOU object 13 g odto.Uprlmt on limits Upper condition. rgy ∼ /.TeXryplei single is pulse X-ray The s/s. inCmr nthe on Camera tion 2. 14 esr aac ewe the between balance ressure ′ -a pulsar x-ray WM en httecs is case the that means FWHM) ′ ihRslto aea(HRC) Camera Resolution High y30 by − 91:0(20) hr san is There (J2000). 59:16:20 L 3 X n David and , ′ niiulpoosare photons Individual . 7 = − ′′ . 960 lal iden- clearly 591620, 2 1 R ies of pixels) HRC (15 1.8 × − Chandra ′′ 10 µ l 591620 16802 × ;w corrected we s; sics.rutgers.edu N 32 6 H ′′ erg = oto of portion 2 Hughes & Slane

18'' 36.91 HRC Data 36.91 Model 36.91 Difference (smoothed)

33.03 33.03 33.03

29.14 29.14 29.14

25.26 25.26 25.26

20'' 21.38 21.38 21.38

17.49 17.49 17.49

13.61 13.61 13.61

9.72 9.72 9.72

-59o 16' 22'' 5.84 5.84 5.84

1.96 1.96 1.96

0.00 0.00 0.00

4.7 10.6 16.6 22.5 28.5 34.4 40.4 4.7 10.6 16.6 22.5 28.5 34.4 40.4 4.7 10.6 16.6 22.5 28.5 34.4 40.4 11h 24m 39.3s 39.0s 11h 24m 39.3s 39.0s 11h 24m 39.3s 39.0s

Fig. 1.— A portion of the Chandra high resolution camera image centered on CXOU J112439.1−591620. In the left and middle panels the grayscale is linear from 0 to 15 HRC counts per pixel (0.1318′′ square). In the right panel the linear grayscale extends from -1.4 to 1.8 counts per pixel. This last image was smoothed with a 1 pixel σ gaussian kernel. Coordinates are given in epoch J2000.

to 221, 222, 223, and 224 temporal bins. A coherent FFT X-ray and radio pulse phases and, due to apparent rota- of the entire light curve showed no statistically significant tional instabilities in the neutron star, it is not possible pulsed signal for any of these cases. The distribution of to extrapolate the radio ephemeris from September 2001 Fourier powers was consistent with noise and the individ- back to July 2001 accurately enough to measure relative ual peak Fourier powers obtained were 30.4, 31.9, 32.9, and phases. 34.2 for the four cases, respectively, none of which are sta- As an additional check on the detection of pulsed X- tistically significant. As a verification of our methods and ray emission, we applied the last search iteration to the IDL software we applied the same programs to the HRC first and second halves of the data set (split in time) in- data of PSR B0540−69.3, observed on 22 June 2000 (Ob- dependently. The pulse was detected in each half at the 2 sID 1745) using the same configuration as our data. The appropriate Z2 value and pulsation frequency and with pulsar was easily detected at a frequency of 19.7941 Hz similar light curve shapes. with a peak Fourier power of 50.3 (99.998% significance). A much more sensitive search for X-ray pulsations is possible by narrowing the range of trial frequencies to be consistent with the radio pulse and a reasonable range of ˙ 2 P values. We employed the Zn test (Buccheri et al. 1983) which applies a harmonic analysis to the phases of photon arrival times for a given trial pulsation frequency. One advantage of the method, compared to epoch-folding for example, is that it requires no binning. Another is that, even for as few as 100 detected photons, the statistic is distributed like χ2 with 2n degrees of freedom. In our searches we use n = 2. We searched eleven trial frequencies spaced by ∆f = 1×10−5 Hz (roughly the frequency resolution of our data) and centered on the expected value based on extrapolating the radio ephemeris to the midpoint of the HRC observa- Fig. 2.— Pulse phase light curve for PSR J1124−5916 folded 2 modulu the best-fit period of 0.13530915 s. Two complete periods tion (MJD = 52105.18). The peak Z2 value was 22.6 cor- are shown. Note the suppressed zero on the y-axis. Also plotted responding to the 99.8% significance level. The search was are the Fourier series estimator (de Jager, Swanepoel, & Rauben- refined by reducing ∆f to 2 × 10−6 Hz and again search- heimer 1986) of the light curve (solid curve) and its 1 σ uncertainty ing eleven trial frequencies, this time centered on the most (dashed curves). likely previous pulsation frequency. This iteration yielded The pulse in the X-ray band is single peaked and sym- 2 a peak Z2 value of 27.9 (99.97% significant or approxi- metric (see Fig. 2), similar to the radio pulse, although the mately 3.6 σ) at a period of 0.13530915 s. The period X-ray pulse width (FWHM ∼ 0.23P ∼ 83◦) is somewhat − error (4 × 10 8 s, 1 σ) was determined using a bootstrap broader than the radio one. The smooth curve in figure 2 is algorithm. In Table 1 we quote observed properties of the a Fourier series estimate (de Jager, Swanepoel, & Rauben- X-ray pulsar. By comparing our pulse period to the value heimer 1986) of the light curve employing two harmonics. obtained by Camilo et al. (2002) roughly two months later If we assume the pulse extends over phase bins 0.43–0.90, ′′ we derive a period derivative of P˙ =7.62±0.06×10−13 s/s we determine the fraction of pulsed X-rays in the 2 radius that differs by ∼2.5 σ from the value quoted in the radio extraction region to be 11 ± 1%. This includes contribu- discovery paper. At present we do not know the relative tion from the diffuse compact nebula, which we quantify next. X-Ray Pulsar in SNR G292.0+1.8 3

˙ 2 TABLE 1 to balance the ram pressure of the wind, Pw = E/4πcrw, Properties of X-ray PSR J1124−5916 assuming spherical symmetry. The mean radius of the compact nebula is 0.036 d6 pc and the spin down energy Parameter Value loss of the pulsar is E˙ = 1.2 × 1037 erg s−1 (Camilo et × −9 −2 −3 al. 2002), so the pressure is Pw =2.6 10 d6 erg cm . R.A. (J2000) 11 24 39.1 We estimate the pressure in the PWN from the prop- Decl. (J2000) −59 16 20 erties of the radio emission, which we take from Gaensler Period, P (s) 0.13530915(4) & Wallace (2003), and the theory of synchrotron emis- Epoch (MJD) 52105.18 sion (Longair 1994). Under the minimum energy con- Observation span (hr) 14.3 ∼ × −10 −4/7 dition we find that PPWN,min 1.3 10 d6 erg FWHM of pulse ∼0.23P cm−3 assuming equal energy densities in the protons and + 9 Pulsed fraction (%) 91−24 electrons and a volume filling factor of unity. The av- HRC rate (s−1) 0.0032(8) erage nebular magnetic field under these conditions is −2 7 −2 21 ∼ / Column density, NH (cm ) 3.1(4) × 10 Bmin 48 d6 µG, which implies a very short syn- ∼ 3/7 −1/2 Photon index, Γ 1.6(1) chrotron lifetime, t 140 d6 (hν/2 keV) yr, for the −1 32 2 Luminosity, LX (0.3–10 keV) (erg s ) 7.2 × 10 (D/6 kpc) electrons giving rise to the X-ray emission. Such a short (Unabsorbed) lifetime is inconsistent with the observation that the X-ray synchrotron nebula covers as large an extent as the radio nebula does. Note.—Numbers in parentheses represent 1 σ uncertainties in A possible solution to these discrepancies lies in relaxing the least significant digits quoted. the minimum energy condition. If we move in the direc- 3. extended compact nebula tion of a smaller mean nebular magnetic field we resolve the lifetime issue. A magnetic field strength of ∼< 8µG Shown in the middle panel of figure 1 is our best fit spa- would ensure that the X-ray synchrotron cooling time is tial model for the HRC data: an unresolved point source ∼> 2000 yr. In order that the pressure in the synchrotron (i.e., a gaussian whose best-fit angular size is consistent nebula be sufficiently strong to balance the ram pressure of with the Chandra PSF) and an elliptical gaussian with a ∼ ′′ the pulsar’s wind requires a value of B 3 µG. The total FWHM of 1.8 (along the major axis) and an axial ratio of energy in the nebula would then be ∼4 × 1049 ergs, con- 2. In this model the point source contains 160 ± 40 X-ray tained nearly entirely in particles. Since this energy has events while the extended elliptical component contains come from the spin-down of the pulsar, it sets a constraint 1440 events. Compared to the number of pulsed events we on the initial spin period: P0 ∼ 22 ms for a canonical NS ± 45 2 detect (146 13), it is clear that the point source itself is momentum of inertia of I ≡ 10 g cm . This P0 value highly pulsed with a pulsed fraction of >65% in the HRC is considerably less than the value of ∼90 ms estimated band. by Camilo et al. (2002). The simplest way to accommo- The rightmost panel in Figure 1 shows the difference be- date our low value for the initial spin period would be to tween the HRC data and the best-fit image model. There increase the true age of the pulsar to ∼2800 yr or more. is good evidence for excess X-ray emission above that It is important to note that neither the magnetic field given by the model, to either side of the point source nor the pressure is expected to be uniform in PWNe, as and oriented generally in the SE-NW direction. One pos- we assumed in the calculations above. In the Kennel & sibility is that the excess emission comes from a pair of Coroniti (1984) model for the Crab Nebula the total pres- jets. This feature is nearly aligned with the direction from sure is greatest at the termination shock and then falls the current position of the point source back toward the by factors of 3–10 at larger radii. In addition, equipar- center of the SNR (toward the NW), which would indi- tition between particles and fields is attained only at a cate aligned spin axis and proper motion directions for significant distance from the pulsar; near the termination PSR J1124−5916, as seen in the Crab and Vela pulsars. shock the magnetic field is low and the pressure is particle- On the other hand, it is also possible that the excess emis- dominated. Because of the higher central pressure, we ex- sion arises from a toroidal structure in the nebula (like pect the volume-averaged magnetic field under this model the torus in the Crab Nebula) seen in projection. In this to be somewhat larger than that estimated above, which scenario the torus would be nearly aligned with the major would have the effect of relaxing the energetics constraint axis of the compact nebula. By analogy to the Crab and on the pulsar’s initial spin period. Our apparent need for its pulsar, we would therefore expect that the spin axis to a particle-dominated PWN in G292.0+1.8 is suggestive of be perpendicular to the long axis of the compact nebula a low value for the magnetization parameter in the context (i.e., aligned NE-SW). This would put the pulsar’s spin of this model. We note, however, that interaction between axis nearly perpendicular to its proper motion direction. the reverse shock and the PWN (which has not yet been The current Chandra data do not allow us to discriminate conclusively established) may offer an alternate explana- between these possibilities. tion for why the nebula is far from the minimum energy The extended compact nebula is the only emission fea- condition. Further study of these issues, although beyond ture in the PWN within an arcmin or so of the pulsar the scope of our work here, is clearly warranted. and therefore is the only plausible candidate for the pul- sar wind termination shock. If we interpret the edge of the nebula with the location of this shock, we can then esti- neutron star cooling mate the confining pressure (i.e., in the PWN) necessary 4. 4 Hughes & Slane

The NS in G292.0+1.8 is quite young with a most likely from the nebular spectrum) to a constraint on the black- age range of ∼2000 yrs to ∼2900 yrs, corresponding to body normalization as a function of TBB. This constraint, the free expansion age of the O-rich knots and the pulsar which is fully consistent with the one from the ACIS-S characteristic age, respectively. According to NS cooling spectral analysis, is shown as the thin curve in figure 3. models (e.g., Tsuruta 1998; Page 1998), the surface tem- Recent work (Gaensler & Wallace 2003) suggests that perature at this age should be high enough to produce de- the distance to G292.0+1.8 is ∼6 kpc. Using this value tectable X-ray emission. As shown by Hughes et al. (2001), and assuming a 12 km radius for the NS, we obtain a con- 6 the ACIS-S spectrum of the pulsar is fully consistent with straint of TBB < 1.18×10 K on the surface temperature of a single absorbed power-law. Here we determine the upper the NS. The expected temperature, assuming standard NS limit to the intensity of an additional blackbody spectral cooling models, is 1.28×106 K (Page 1998). Although this component as a function of its temperature, TBB. We is suggestive of the presence of exotic cooling processes, utilized two independent spectra extracted from the CTI- systematic uncertainties make this result less secure than corrected data (Park et al. 2002): one from a 3×3 pixel the recent result on the apparent need for exotic cooling (1.5′′×1.5′′) region centered on the pulsar, and another, processes for the NS in 3C 58 (Slane, Helfand, & Murray comprising the diffuse nebula, from an ellipse of size 7×11 2002). The NS in G292.0+1.8 would be consistent with pixels (3.4′′× 5.4′′) excluding the central pulsar region. standard cooling if it were as distant as 7 kpc, or if the The pulsar spectrum was fit to the sum of a blackbody compact star’s radius were as small as 10 km. On the other and a power-law model including absorption, while the hand pure blackbody spectral models tend to overpredict nebular spectrum was fit to an absorbed power-law model (by factors of 1.5 or more) the effective temperature of alone. This latter spectrum served as an independent con- NS surfaces when light element atmospheres are included straint on the column density, which was constrained to (Lloyd, Hernquist, & Heyl 2002). be the same between the two spectra. For reference, table 1 lists pure power-law spectral parameters for the pulsar. 5. limits on an optical counterpart For a given fixed value of T , the ACIS-S data set an BB Optical emission from isolated pulsars within supernova upper limit on the allowed normalization (or flux) of the remnants has currently been detected from only four ob- blackbody component. One can express the normalization − limit in terms of the square of the ratio of the blackbody jects: PSR B0531+21 (Crab), PSR B0540 69.3 (in the LMC), PSR B1509−58 (G320.4−1.2), and PSR B0833−45 emitter’s radius to its distance. The 3 σ limit on this ratio (Vela) (see, for example, Nasuti et al. 1997 and references as a function of T is plotted in figure 3. BB therein). The first three are very young pulsars (1000– 2000 years old), while the pulsar in Vela is considerably older (∼10,000 yrs), although it is still rather young com- pared to the average radio pulsar. Across the optical band these pulsars show flat power-law spectra (α ∼ 0 for −α Fν ∝ ν ), although their intrinsic luminosity densities 2 (i.e., Lν =4πD Fν ) span 5 orders of magnitude from 0.5– 2 × 1019 ergs−1 Hz−1 (PSR B0531+21, PSR B0540−69.3, and PSR B1509−58) to 3–6 × 1014 ergs−1 Hz−1 (PSR B0833−45). In terms of age and remnant optical prop- erties (i.e., the presence of high velocity, oxygen-rich op- tical emission), G292.0+1.8 most closely resembles SNR 0540−69.3. However in terms of spin-down energy loss (∼1037 erg s−1), the pulsar in G292.0+1.8 is more similar to PSR B0833−45 and PSR B1509−58. With Chandra we have localized the G292.0+1.8 pulsar Fig. 3.— Constraint on the normalization of a blackbody spec- to an absolute position accuracy of ∼1′′. Within double tral component vs. its temperature from fits to the time-averaged ACIS-S spectrum of PSR J1124−5916 (thick solid curve). The thin this error circle there is no optical counterpart visible in solid curve indicates the constraint based on the unpulsed HRC the Digitized Sky Survey. We have obtained an upper limit count rate of the X-ray pulsar. The allowed region lies to the left on optical emission from the pulsar, B & 22, based on a and below the curves shown. The dashed lines show the tempera- narrow-band blue continuum image of G292.0+1.8 taken ture constraint for the nominal value of distance to G292.0+1.8(6 kpc) and a 12 km radius NS. The vertical dotted line shows the by P.F. Winkler and K.S. Long from the CTIO 4-m in temperature expected for a standard NS cooling curve. 1991. This corresponds to an intrinsic luminosity density 18 −1 −1 of Lν < 3 × 10 ergs Hz (assuming a distance of 6 ∼ Since the ACIS-S spectrum is consistent with an en- kpc and extinction of AB 2.3). This is about an order tirely nonthermal origin, the pulsed emission seen in the of magnitude less than the optical emission of the Crab − HRC, which comprises >65% of the total HRC rate from and SNR 0540 69.3 pulsars, but is only about a factor − the pulsar, therefore must be dominated by nonthermal, of two less than the optical emission from PSR B1509 58 i.e., magnetospheric, emission as well. The unpulsed HRC (Caraveo, Mereghetti, & Bignami 1994). A considerably emission, however, can be used to set another constraint on fainter upper limit, based on data acquired at CTIO in the mean surface temperature of the NS. We convert the 3 April 2002, will be the subject of a forthcoming article. σ upper limit on the unpulsed HRC count rate (2.8×10−3 s−1), assuming the 3 σ upper limit on the column den- We are grateful to Fernando Camilo and Bryan Gaensler 21 −2 sity to the pulsar (NH = 4.75 × 10 atom cm , derived for sharing results or data prior to publication and X-Ray Pulsar in SNR G292.0+1.8 5 to Frank Winkler for supplying the optical image of time. We also thank Simon Johnston for his useful com- G292.0+1.8. Mike Juda gave us some helpful advice re- ments as referee. Partial support for this research was garding the HRC data. We thank Karen Lewis and John provided by Chandra grant GO1-2052X to JPH. Nousek for their help with the initial proposal for HRC

REFERENCES

Buccheri, R., et al. 1983, A&A, 128, 245 Longair, M. S. 1994, High Energy Astrophysics, Vol 2, 2nd edition Camilo, F., Manchester, R. N., Gaensler, B. M., Lorimer, D. R., (Cambridge: Cambridge University Press) p. 292ff Sarkissian, J. 2002, ApJ, 567, L71 Murdin, P., & Clark, D. H. 1979, MNRAS, 189, 501 Caraveo, P. A., Mereghetti, S., & Bignami, G. F. 1994, ApJ, 423, Nasuti, F. P., Mignani, R., Caraveo, P. A., & Bignami, G. F. 1997, L125 A&A, 323, 839 De Jager, O. C., Swanepoel, J. W. H., & Raubenheimer, B. C. 1986, Page, D. 1998, in The Many Faces of Neutron Stars, ed. R. Buccheri, A&A, 170, 187 J. van Paradijs, & M. A. Alpar (Dordrecht: Kluwer), 539 Gaensler, B. M., & Wallace, B. J. 2003, ApJ, in press (astro- Park, S., Roming, P. W. A., Hughes, J. P., Slane, P. O., Burrows, ph/0305168) D. N., Garmire, G. P., & Nousek, J. A. 2002, ApJL, 564, L39 Hughes, J. P., Slane, P. O., Burrows, D. N., Garmire, G. P., Nousek, Rees, M. J., & Gunn, J. E. 1974, MNRAS, 167, 1 J. A., Olbert, C. M., & Keohane, J. W. 2001, ApJ, 559, L153 Slane, P. O., Helfand, D. J., & Murray, S. S. 2002, ApJ, 571, L45 Kennel, C. F., & Coroniti, F. V. 1984, ApJ, 283, 694 Tsuruta 1998, Rev. Mod. Phys.