Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://hea-www.harvard.edu/~kim/pap/PKS0637_2000.pdf Äàòà èçìåíåíèÿ: Mon Dec 6 20:20:39 2010 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 00:35:50 2012 Êîäèðîâêà: koi8-r Ïîèñêîâûå ñëîâà: cygnus

THE ASTROPHYSICAL JOURNAL, 542 : 655 õ 666, 2000 October 20
( 2000. The American Astronomical Society. All rights reserved. Printed in U.S.A.

THE CHANDRA X-RAY OBSERV AT ORY RESOLVES THE X-RAY MORPHOLOGY AND SPECTRA OF A JET IN PKS 0637[752 G. CHARTAS,1 D. M. WORRALL,2,3 M. BIRKINSHAW,2,3 M. CRESITELLO-DITTMAR,3 W. CUI,4 K. K. GHOSH,5 D. E. HARRIS,3 E. J. HOOPER,3 D. L. JAUNCEY,6 D.-W. KIM,3 J. LOVELL,6 S. MATHUR,7 D. A. SCHWARTZ,3 S. J. TINGAY,8 S. N. VIRANI,3 AND B. J. WILKES3
Received 2000 March 10 ; accepted 2000 May 11

ABSTRACT The core-dominated radio-loud quasar PKS 0637[752 (z \ 0.654) was the ïrst celestial object observed with the Chandra X-Ray Observatory, oering the early surprise of the detection of a remarkable X-ray jet. Several observations with a variety of detector conïgurations contribute to a total exposure time with the Chandra ACIS of about 100 ks. A spatial analysis of all the available X-ray data, making use of Chandraîs spatial resolving power of about 0A , reveals a jet that extends about 10A to the .4 west of the nucleus. At least four X-ray knots are resolved along the jet, which contains about 5% of the overall X-ray luminosity of the source. Previous observations of PKS 0637[752 in the radio band had identiïed a kiloparsec-scale radio jet extending to the west of the quasar. The X-ray and radio jets are similar in shape, intensity distribution, and angular structure out to about 9A, after which the X-ray brightness decreases more rapidly and the radio jet turns abruptly to the north. The X-ray luminosity of the total source is log L B 45.8 ergs s~1 (2õ10 keV) and appears not to have changed since it was X observed with ASCA in 1996 November. We present the results of ïtting a variety of emission models to the observed spectral distribution, comment on the nonexistence of emission lines recently reported in the ASCA observations of PKS 0637[752, and brieÿy discuss plausible X-ray emission mechanisms. Subject headings : galaxies : active õ galaxies : jets õ quasars : individual (PKS 0637[752) õ X-rays : galaxies
1

. INTRODUCTION

Until recently, most of our knowledge regarding the spatial structure and spectral shape of extragalactic jets has relied on observations performed in the radio and optical bands. X-ray detections and upper limits have provided insights into the physical conditions responsible for the observed radiation from knots and hot spots in extragalactic radio jets. However, in most cases the poor spectral and spatial resolution available has made the interpretation of the X-ray data difficult. The spectral energy distributions (SEDs) seen from jets have been ascribed to combinations of synchrotron radiation, synchrotron self-Compton (SSC) radiation, and thermal bremsstrahlung from shock-heated gas near the jets. In particular, the synchrotron and SSC models have been successfully employed to explain the X-ray emission from hot spots and jets in M87 (Biretta, Stern, & Harris 1991), Cygnus A (Harris, Carilli, & Perley 1994), and 3C 295 (Harris et al. 2000). In the cases of 3C 273 (Harris & Stern 1987) and Pictor A (Meisenheimer et al. 1989), none of the standard processes yield satisfactory results.
1 Astronomy and Astrophysics Department, Pennsylvania State University, University Park, PA 16802 ; chartas=astro.psu.edu. 2 Department of Physics, University of Bristol, England. 3 Harvard-Smithsonian Center For Astrophysics, Cambridge, MA 02138. 4 MIT Center for Space Research, 70 Vassar Street, Cambridge, MA 02139. 5 Space Sciences Laboratory, NASA/Marshall Space Flight Center, Mail Code ES84, Huntsville, AL 35812. 6 Australia Telescope National Facility, P.O. Box 76, Epping, NSW 2121, Australia. 7 Ohio State University. 8 Jet Propulsion Laboratory, California Institute of Technology, Mail Stop 238-332, 4800 Oak Grove Drive, Pasadena, CA 91109.

The Chandra X-Ray Observatory, launched on 1999 July 23, provides a signiïcant improvement over previous missions in combined spatial and spectral resolution (see Weisskopf & OîDell 1997 ; van Speybroeck et al. 1997), which we expect will result in a signiïcant increase in the number of detected and resolved X-ray jets. PKS 0637[752 is the ïrst X-ray jet to have been discovered by Chandra. The quasar was originally detected in X-rays with the Einstein Observatory (Elvis & Fabbiano 1984). Since then a peculiar emission-line feature at D0.97 keV was claimed in an ASCA SIS observation of the source (Yaqoob et al. 1998). One of two Ginga observations of PKS 0637[752 has been reported as showing a marginal detection of an Fe Ka line with an equivalent width of 103 ^ 85 eV (^1 p errors ; Lawson & Turner 1997). In ° 2 we describe the spectral and spatial analysis of the core and jet and our observations of the radio jet. The properties of sources in the vicinity of PKS 0637[752 are also brieÿy presented. Section 3 contains a brief description of our attempt to apply standard jet models to the observed SED of PKS 0637[752. We provide a thorough investigation of the underlying emission processes in a companion paper (Schwartz et al. 2000), where more complex models are considered. In the following, we assume a Friedman cosmology with H \ 50 km s~1 Mpc~1 and q \ 0. 0 0
2.

DATA REDUCTION AND ANALYSIS

Twenty-ïve observations of PKS 0637[752 were made between 1999 August 14 and August 24 with the Chandra ACIS-S during the orbital activation and checkout phase of the mission. Table 1 summarizes several ACIS conïguration parameters, 50% encircled energy radii, exposure times, and estimated net count rates over the entire ACIS band corresponding to each individual observation. 655


TABLE 1 CHANDRA OBSERVATIONS OF PKS 0637[752

Observation Date (1) 1051 1052 62558 62556 62555 62554 62553 62552 62551 62550 62549 1055 1056 1058 1059 1060 1062 1063 472 473 474 475 476 1034. 1041. 19116. 5068. 5138. 11367. 5232. 5213. 5343. 5469. 6426. 2036. 1757. 1757. 1760. 1760. 1760. 1760. 5809. 4678. 4856. 4856. 4856. 3.24 3.24 3.24 3.24 3.24 3.24 1.541 1.541 1.541 1.541 1.541 3.24 3.24 3.24 3.24 3.24 3.24 3.24 0.941 0.941 0.941 3.24 0.941 0.0 0.0 1.0 0.9 0.65 0.25 0.0 [0.5 [0.25 0.2 0.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.11 [0.09 [0.19 [0.98 2.0 2.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.4 [1.4 [3.0 [1.4 1.4 2.8 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.4 1.4 [2.8 [1.4 1.4 [2.8 0.0 0.0 0.0 0.0 0.0 1.05 0.95 1.00 0.95 0.80 0.55 0.40 0.70 0.40 0.50 0.55 0.85 1.10 1.00 1.00 1.20 1.10 2.00 1.05 0.4 0.4 0.45 1.10 408 391 7196 1799 1554 2664 2098 2373 2108 2266 2777 659 744 710 555 781 758 1019 3417 2206 2255 2386 2967 5 2 63 18 18 40 16 20 18 33 22 8 5 7 5 6 9 15 20 21 21 22 10

Observation ID (2)

Exposure (s) (3)

Frame Time (s) (4)

SIM x (mm) (5)

Src y (arcmin) (6)

Src z (arcmin) (7)

HPR (arcsec) (8)

N Core (counts) (9)

N In (counts) (10)

N Out (counts) (11) 28 28 502 119 155 281 188 148 148 128 152 50 42 41 22 51 54 40 140 117 112 100 113

N Bkg (counts arcsec~2) (12) 0.028 0.045 0.498 0.111 0.108 0.267 0.110 0.100 0.108 0.099 0.133 0.057 0.061 0.044 0.250 0.155 0.097 0.115 0.095 0.085 0.088 0.088 0.167

1999-08-14T10 1999-08-14T11 1999-08-14T13 1999-08-14T19 1999-08-14T20 1999-08-14T22 1999-08-15T02 1999-08-15T04 1999-08-15T05 1999-08-15T07 1999-08-15T09 1999-08-16T06 1999-08-16T06 1999-08-16T17 1999-08-16T18 1999-08-16T19 1999-08-16T20 1999-08-16T21 1999-08-20T02 1999-08-20T04 1999-08-20T06 1999-08-20T07 1999-08-20T09

: : : : : : : : : : : : : : : : : : : : : : :

49 36 53 15 51 24 49 18 55 36 12 05 56 43 29 16 53 35 26 54 17 54 30

: : : : : : : : : : : : : : : : : : : : : : :

39 00 09 24 39 18 29 30 56 04 04 37 36 28 48 07 19 07 38 34 37 18 57

...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ......

NOTES.õCol. (5) : SIM is the distance along the optical axis of the scientiïc instrument module from the best focus location ; col. (6) : Src is the distance along the y-direction from the nominal aim x y point ; y is the direction of grating dispersion ; the ACIS-S array nominal aim point falls on chip S3, 2@ to the right (]y) of the edge of the chip ; col. (7) : Src is the distance along the z-direction from the .0 z nominal aim point ; z is the direction normal to the grating dispersion ; col. (8) : HPR is the half-power radius ; col. (9) : N are the detected events from the core component extracted from circles centered Core on the core with radii of 5A ; col. (10) : N are the detected events from the inner jet component extracted from rectangular regions having the left lower corners set at (x ] 4@@, y [ 1A ) and the right upper .5 In c c corners set at (x ] 6A , y ] 1A5), where x , y are the centroid locations of the core for each observation ; col. (11) : N are the detected events from the outer jet component extracted from circles centered .5 . c c cc Out on the X-ray knot WK8.9 with radii of 2A ; col. (12) : N are the detected background events arcsec~2 extracted from annuli centered on the core with inner and outer radii of 45A and 55A, respectively. .5 Bkg Only events with standard ASCA grades 0, 2, 3, 4, and 6 were extracted.


CHANDRA RESOLVES JET IN PKS 0637[752 The primary purpose of these observations was to focus the ACIS-S detector (in SIM ) and to determine the posix tion of the Chandra mirror optical axis by pointing to PKS 0637[752 at dierent (Src , Src ) osets. (Deïnitions for y z Src , Src , and SIM are presented in the notes of Table 1.) y z x The spectral and spatial analysis of the data is complicated by the numerous conïgurations used during the observation of PKS 0637[752. The point-spread function (PSF) corresponding to each observation diers signiïcantly because of the nature of the calibration being performed. In particular, the 50% encircled energy radii varied between 0A at best focus and about 1A at 3@ o-axis point.4 ing or poor focus (see Fig. 1 and Table 1). The jet (° 2.1) was spatially resolved from the core and made little impact on the determination of the best focus and best optical axis location. A second important property of the data that we have accounted for in our analysis is pileup (see ° 6.17 of the Chandra Observatory Guide).9 Whenever the separation of two or more X-ray photons incident on a CCD is less than a few CCD pixels, and their arrival time lies within the same CCD frame readout, the CCD electronics may regard them as a single event with an amplitude given by the sum of the electron charge in the 3 ] 3 neighborhood of the pixel with the maximum detected charge. A detected CCD event is characterized by the total charge within the 3 ] 3 island and the distribution of the charge (often referred to as the grade of the event) within that island. Pileup may alter the grades and charges of events, thus aecting both their spatial and spectral distribution. A manifestation of pileup in observed spectra may be a reduction of detected events, spectral hardening of the continuum component, and the apparent distortion of the PSF of pointlike objects. The presence of pileup is apparent in the observed spectra of the core of PKS 0637[752. In particular, we observed a signiïcant change in count rate of the core component of PKS 0637[752 as a function of focus position (SIM ) x along the optical axis. As shown in Figure 1, the apparent

657

count rate is lowest at the best focus position and increases as the scientiïc instrument module (SIM) is moved away from this location. A parabolic function was used to ït the count rates and encircled energy radii as a function of focus position. We estimate a SIM focus location of D[0.7 mm x corresponding to the minimum of both count rate and encircled energy radii. Count rates for a similar test performed with a shorter CCD frame readout time showed a similar behavior but with detected count rates systematically larger than those for the standard 3.24 s full-frame readout time. The observed count rate variation is due to the pileup eect and not intrinsic to the quasar. As the telescope approaches best focus, most photons fall onto a single CCD pixel, and the enhanced pileup leads to a decrease in detected count rate. Correcting a CCD observation of an X-ray source for pileup is quite complicated. The dependence of pileup on several eects such as the incident X-ray ÿux, the spectral and spatial distribution of the events, the grade selection scheme used, and the detection cell size adopted make the problem of restoring to the unpiled up spectrum nontrivial. Fortunately, the jet of PKS 0637[752 is not aected by pileup since the jet is a relatively low count rate extended source. 2.1. X-Ray Morphology To obtain a high signal-to-noise image of PKS 0637[752, we made use of all observations with observed half-power radius (HPR) less than 1A . For each obser.2 vation, a PSF appropriate for the focus and aim point position was created employing the simulation tool MARX v2.2 (Wise et al. 1997).10 The input spectrum assumed in the PSF simulations was that derived from the best-ït Chandra spectrum of the outer jet of PKS 0637[752 (see ° 2.3). Speciïcally, we used an absorbed power law with a column density of N \ 11.8 ] 1020 cm~2 and a photon index of 1.83. We H examined the sensitivity of the resulting deconvolved image to the spectral slope used in the PSF simulations and found no particular spectral dependence for photon indices ranging between 1.7 and 2.3. The simulated PSFs were binned to a subpixel scale of 0A125. To avoid aliasing eects, the Chandra X-Ray Center . (CXC) processing incorporates a randomization of each position by ^0A246 (1 ACIS pixel \ 0A492). Residual errors . . in the aspect solution are expected to add a "" blurring îî to detected photon positions of D0A rms in diameter. To .3 simulate aspect errors and position randomization, we convolved each generated PSF with a Gaussian with p \ 0A25. . The X-ray photon event positions for each observation were also binned to 0A125, and the resulting X-ray image is . shown in Figure 2 (top panel). A maximum likelihood deconvolution technique, using the appropriate simulated PSF, was applied to each individual observation. The resulting deconvolved images were aligned on the core centroid and combined to produce the total deconvolved image of PKS 0637[752 shown in Figure 2 (middle panel). The eective resolution of the deconvolved X-ray image was estimated by deconvolving the simulated PSFs of each observation, stacking the simulated PSFs, and determining the FWHM of the deconvolved stacked PSF. Our analysis yields an eective resolution for the stacked deconvolved . image of D0A37. The PSFs simulated for this analysis are
10 http ://space.mit.edu/ASC/MARX/.

FIG. 1.õ50% and 90% encircled energy and observed count rate as a function of SIM position. The increased PSF half-power radius (upper panel) and count x rate (lower panel) with SIM position away from best focus increases because pileup decreases.

9 http ://asc.harvard.edu/udocs/docs/POG/MOPG/index.html.


658

CHARTAS ET AL.

Vol. 542

FIG. 2.õT op panel : X-ray image of PKS 0637[752 created by stacking all observations with half-power radii less than 1A . Middle panel : Total .2 maximum likelihood deconvolved X-ray image of PKS 0637[752 produced by stacking all deconvolved images of observations with half-power radii less than 1A . L ower panel : ATCA 8.6 GHz image of PKS 0637[752 restored with a circular beam of 0A 4 FWHM. .2 .8

not appropriate for the piled up region of the core. This is why any deconvolved structure within D2A from the core should not be considered real. The jet region beyond D2A does not suer from pileup eects, and the simulated PSFs are appropriate for deconvolving the jet region. Several interesting structures have become more apparent in Figure 2 (middle panel). A well-collimated X-ray jet is seen to originate in the core and to extend approximately 10A to the west, and within the jet at least four knots are clearly resolved.

2.2. Radio Morphology The radio jet was imaged at 4.8 and 8.6 GHz on 1999 August 19 and on 1999 September 21 using the Australia Telescope Compact Array (ATCA), which has a similar arcsecond resolution at 8.6 GHz as Chandra and so provides a powerful structural comparison. The resulting 8.6 GHz radio image (Schwartz et al. 2000 ; Lovell et al. 2000) is shown in Figure 2 (lower panel), where the coincidence with the X-ray knots is apparent. The three farthest X-ray knots,


No. 2, 2000

CHANDRA RESOLVES JET IN PKS 0637[752

659

WK7.8, WK8.9, and WK9.7, appear to be embedded in extended and diuse X-ray emission. The X-ray jet appears to bend southwest after its encounter with the ïrst X-ray knot, WK5.7, and then bend northwest after the encounter with the fourth X-ray knot WK9.7. The radio image of the PKS 0637[752 jet (Fig. 2, lower panel) indicates that the radio jet also bends in the northwest direction after its encounter with WK9.7. The X-ray emission after knot WK9.7 drops substantially to become undetectable after several arcseconds. A comparison between the X-ray and radio intensity proïle along the jet (Fig. 3) shows close alignment of the X-ray and radio jets out to 10A and relatively well-matched X-ray and radio knots at 5A , 7A , 8A , .7 .8 .9 and 9A from the core. The radio proïle in Figure 3 was .7 produced from the 8.6 GHz image (Schwartz et al. 2000 ; Lovell et al. 2000). The position angles of the radio and X-ray jets appear to be similar, as shown in Figure 4, where we have used the deconvolved X-ray image and the same 8.6 GHz radio map as before to plot the position angle of the ridges of peak radio and X-ray brightness along the jet with respect to the core of PKS 0637[752. 2.3. Spectral Analysis of the Core Component To determine the continuum spectral shape of the core component of PKS 0637[752, we considered only the two least piled up on-axis observations corresponding to observations 472 and 476. The expected percent loss of counts due to pileup, based on our simulations, is D12%. X-ray events were extracted within a circle centered on the core of PKS 0637[752 with a radius of 5A, and events with standard ASCA grades 0, 2, 3, 4, and 6 only were selected. The background was determined by extracting events within annuli centered on the core with inner and outer radii of 15A and 20A, respectively. Spectra were binned to have a minimum of 20 counts per bin such that s2 statistics can be used without low count corrections. Spectral ïts were performed using the standard software tool XSPEC (Arnaud 1996), and we compared these results

FIG. 4.õX-ray and radio position angle of the ridge of peak brightness along the jet with respect to the core of PKS 0637[752. The radio position angle plot was derived from the 8.6 GHz image (Schwartz et al. 2000 ; Lovell et al. 2000).

FIG. 3.õX-ray and radio intensity proïle of jet along right ascension direction integrated ^1A perpendicular to the jet. The X-ray proïle provides counts in 0A125 increments, and the radio proïle provides the 8.6 . GHz ÿux density beam~1 in 0A bins. The X-ray proïle was produced from .1 the deconvolved X-ray image, and the radio proïle was produced from the 8.6 GHz image (Schwartz et al. 2000 ; Lovell et al. 2000). The 8.6 GHz beam width is D1A FWHM, and the eective resolution (FWHM) of the X-ray image after deconvolution is D0A . .4

to those provided by the simulator-based spectral ïtting tool LYNX (Chartas et al. 2000, in preparation). A brief description of the LYNX ïtting tool is provided in the Appendix. The telescope pointings are intentionally dithered for the observations of PKS 0637[752 in a Lissajous pattern with an amplitude of about 16A such that smallscale nonuniformities in the CCD quantum efficiency are averaged out. The selected aim point of the telescope for most of the PKS 0637[752 observations was set within 10A from the boundary between ampliïer nodes 0 and 1 of the S3 CCD of ACIS, resulting in the source being dithered across the two nodes. To account for the dierence in ampliïer gains between nodes 0 and 1, events detected from the core component of PKS 0637[752 were separated into two spectra containing events detected from a single node, either 0 or 1. For ïtting the spectra with XSPEC, we used the appropriate response and ancillary ïles provided by the CXC. All errors on best-ït spectral parameters quoted in this paper are at the 90% conïdence level unless mentioned otherwise. Speciïcally, we used the Chandra Interactive Analysis of Observations (CIAO) v1.1 tools MKRMF and MKARF to generate response and ancillary ïles. The focal plane temperature was [100¡C for these observations. The exposure times of each observation are listed in Table 1. A simple power-law plus cold, neutral absorber model was used to ït the core spectrum of PKS 0637[752. The best-ït parameters of the spectral ïts are shown in Table 2. Fit 2 of Table 2 was performed in the energy range 0.2õ 6.0 keV. The reduced s2 of 1.1 for 235 degrees of freedom is l acceptable in a statistical sense. However, the best-ït value of (3.5 ^ 0.5) ] 1020 cm~2 for the column density is signiïcantly below the previously observed ASCA value of (9 ^ 3) ] 1020 cm~2 and the Galactic column density of 9.1 ] 1020 cm~2 (Dickey & Lockman 1990). In ït 3 of Table 2, we ïxed the neutral column density to the Galactic value and obtained a reduced s2 of 1.7 for 236 degrees of l freedom. We searched for systematic errors in the spectral ïts of the core by repeating ït 3 of Table 2 allowing for the lower bound of the ïtted energy range to vary between 0.2 and 1.5


660

CHARTAS ET AL.
TABLE 2

Vol. 542

MODEL PARAMETERS DETERMINED FROM SPECTRAL FITS TO THE CHANDRA ACIS-S SPECTRA OF THE CORE COMPONENT OF PKS 0637[752 N H (z \ 0) (1020 cm~2) Fluxb (10~12 ergs s~1 cm~2) Lc X (1045 ergs s~1)

Fita 1 2 3 4 ...... ...... ...... ......

Range (keV) 0.2õ 0.2õ 0.2õ 1.0õ 6.0 6.0 6.0 6.0

!

Flux Density at 1 keV (10~13 ergs s~1 cm~2 keV~1)

1.77`0.1 11`2.0 1.1 (2.2) 7.4 5.9 (6.4) ~0.1 ~2.0 1.69`0.05 3.5`0.5 1.3 (2.2) 7.8 4.8 (6.0) ~0.05 ~0.5 2.00`0.04 9.0 (ïxed) 1.3 (1.7) 8.5 7.8 (5.5) ~0.04 1.76`0.08 9.0 (ïxed) 1.1 (2.2) 7.5 5.6 (6.1) ~0.08 a All ïts incorporate a power-law spectrum plus absorption due to cold material at solar abundances. Spectral ït 1 was performed with the LYNX spectral ïtting tool with events extracted from node 1 only. Spectral ïts 2, 3, and 4 were performed with the XSPEC tool with events extracted from nodes 0 and 1. b Fluxes calculated in the ranges 0.2õ2 keV and 2õ10 keV (quoted in parentheses). X-ray ÿuxes are not corrected for Galactic absorption. c Luminosities calculated in the ranges 0.2õ2 keV and 2õ10 keV (quoted in parentheses). Luminosities are in the rest frame and are corrected for Galactic absorption.

s2/(dof ) l 1.4 (86) 1.1 (235) 1.7 (236) 1.1 (128)

keV. We ïnd that the best-ït XSPEC value for the spectral slope varies between 2.00 ^ 0.04 and 1.8 ^ 0.05 for the lower bounds of the ït ranging between 0.2 and 0.8 keV, respectively, and remains constant at about 1.8 for the lower bounds of the ït ranging between 0.8 and 1.5 keV. Plausible explanations for the variation of the spectral slope with the lower bound of the ït include pileup eects of the spectrum of the core and/or uncertainties in the available response and ancillary ïles at energies below D0.5 keV. We therefore restricted the XSPEC ïts of the core component to energies above 1 keV with the neutral column density parameter held ïxed at the Galactic value of N \ 9 ] 1020 H cm~2. The ït within the 1.õ 6 keV range yields a photon index of 1.76 ^ 0.1 (ït 4 in Table 2). Best-ït parameters to the core component were also obtained utilizing the simulator based tool, LYNX (ït 1 in Table 2), and restricting the energy range of the ït between 0.2 and 6.0 keV. This ït yields a column density of (11 ^ 2) ] 1020 cm~2 that is consistent with the Galactic

value and a photon index of 1.77 ^ 0.1 that is consistent with Ginga and ASCA results (Lawson et al. 1992 ; Lawson & Turner 1997 ; Yaqoob et al. 1998). All X-ray photon indices, !, and energy indices, a , follow the convention of E ÿux density Pl~(!~1), where ÿux density is in units of ergs cm~2 s~1 Hz~1 and ! \ a ] 1. E The dierence between the LYNX and XSPEC ït results for the spectrum of the core of PKS 0637[752 can be partially attributed to the presence of pileup in the spectrum of the core, which is not modeled in the XSPEC ïts. In Figure 5, we present the spectrum of the core component of PKS 0637[752 corresponding to observation 476, together with the best-ït model (ït 1 from Table 2) and the ratio of observed spectrum to model. The estimated 2õ10 keV ÿux of 2.2 ] 10~12 ergs s~1 cm~2 (from ït 1 of Table 2) indicates that the core ÿux has not varied, within the quoted error bars, since the 1996 November ASCA observations, especially when one considers that the ASCA observed ÿux corresponds to all emission within a 3@ radius

FIG. 5.õSpectrum from observation 476 of core component of PKS 0637[752 with best-ït model (ït 1 from Table 2)


No. 2, 2000

CHANDRA RESOLVES JET IN PKS 0637[752

661

of the core. X-ray ÿuxes quoted in this paper are not corrected for Galactic absorption. We chose to improve the signal-to-noise ratio for the detection of faint emission lines in the spectrum of PKS 0637[752 by stacking all the available spectra. In the summed spectrum, we expect that pileup will distort the continuum shape and cause emission lines to appear with overtone "" ghosts îî at higher energies. We initially produced a model that best ït the continuum component of the stacked spectrum. The region in the vicinity of the mirror Ir edges (see Table 3) was excluded from the ït. In Figure 6, we plot the stacked spectrum of the core component of PKS 0637[752 with the best-ït continuum component. We also show the ratio of the observedõtoõbest-ït continuum model. The best-ït model yields a reduced s2 of 0.99 for 147 degrees of freedom (dof ). This ratio suggestsl that no signiïcant emission-line features are present. In particular, we do not detect an emission line in the vicinity of 1 keV and at an equivalent width (EW) D60 eV as suggested by a recent ASCA observation of PKS 0637[752 (Yaqoob et al. 1998). An upper limit of 15 eV on EW, at the 90% conïdence level, can be placed on any line with energies 0.8õ1.2 keV (observed frame) and width of 0õ 0.2 keV. There are several spectral features in observed ACIS spectra that do not originate from astrophysical sources but are produced from various sites within the Chandra/ACIS instrument. In Table 3, we list all known instrumental spectral features and their origins. The signiïcant instrumental features are the Al-L, C-K, N-K, O-K, and Al-K UV/optical blocking ïlter absorption edges ; the N-K, O-K, and Si-K CCD absorption edges ; and the Ir-M high-resolution mirror assembly (HRMA ) absorption edges. In the spectra of the core of PKS 0637[752 (see Figs. 5 and 6), the main instrumental spectral features that can be seen are the Ir-M

HRMA absorption edge at 2.085 keV and the O-Ka ACIS absorption edge at 0.536 keV. 2.4. Spectral Analysis of the Inner Jet Component For the purposes of this spectral analysis, we have extracted the spectrum of the jet in a region 4Aõ 6A away .5 from the core and deïne this as the inner jet region. Pileup eects are negligible for the inner jet component because of the extended nature of the emission combined with the relatively low count rate of D4 ] 10~3 counts s~1. The X-ray image of PKS 0637[752 shown in Figure 2 indicates a degree of curvature in the jet that appears to follow the spatial morphology observed in the radio (Schwartz et al. 2000 ; Lovell et al. 2000). The X-ray spectrum of the inner jet component was produced by extracting events from all the observations listed in Table 1. The spatial extraction ïlter chosen was a rectangle with the left lower corner set at (x ] 4@@, y [ 1A ) and the right upper corner set at .5 c c (x ] 6A , y ] 1A5), where x , y are the centroid locations .5 . cc ofcthe core c each observation. In Figure 7, we show the for observed spectrum of the inner jet component with the bestït model assuming a power-law emission process (ït 3 in Table 4). Because of the present uncertainties with response matrices below 0.5 keV for the S3 CCD, the spectral ïts utilizing XSPEC were performed in the 0.6õ 4 keV range, and ïts utilizing the tool LYNX (which appears to be more reliable at energies below 1 keV) were performed in the 0.2õ 4 keV range. Model 1 incorporates a simple absorbed power-law model and yields a photon index of ! \ 2.0 ^ 0.2 (LYNX ; ït 1 in Table 4) or ! \ 2.27 ^ 0.3 (XSPEC ; ït 3 in Table 4). Model 2 incorporates a Raymond-Smith thermal plasma model and a best-ït column density consistent with the Galactic value and

FIG. 6.õStacked spectrum of the core component of PKS 0637[752. No prominent emission line is detected near 1 keV as claimed in a recent ASCA observation of PKS 0637[752.


662

CHARTAS ET AL.
TABLE 3 CHANDRA ACIS INSTRUMENTAL SPECTRAL FEATURES Energy (keV) 0.076 ........... 0.105 ........... 0.107 ........... 0.158 ........... 0.285 ........... 0.402 ........... 0.535 ........... 1.486 ........... 1.559 ........... 1.739 ........... E -E b ......... 0f 1.841 ........... 1.8473 .......... 1.8447 .......... 2.085c .......... 2.112 ........... 2.156c .......... 2.20 ............. 2.410 ........... 2.549c .......... 2.906c .......... 3.183c .......... 7.469 ........... 9.71 ............. 11.44 ........... 11.52 ........... 15.2-E d ...... bias 15.2 ............. Spectral Featurea Al-L absorption edge Si L absorption edge 3 Si L absorption edge 2 Si L absorption edge 1 C Ka absorption edge N Ka absorption edge O Ka absorption edge Al K ÿuorescence line a Al K absorption edge a Si K ÿuorescence line Si K escape peak Si K absorption edge of polysilicon Si K absorption edge of SiO 2 Si K absorption edge of Si N 34 Ir M-V edge Au Ma ÿuorescence line 1,2 Ir M-IV edge Au Mb ÿuorescence line Au Mc ÿuorescence line Ir M-III edge Ir M-II edge Ir M-I edge Ni K ÿuorescence line a Au La ÿuorescence line 1 Au Lb ÿuorescence line 1 Au Lb ÿuorescence line 2 4096 ADUõbias level 4096 ADU

Vol. 542

Origin ACIS-OBF ACIS-CCD ACIS-CCD ACIS-CCD ACIS-OBF ACIS-OBF, ACIS-CCD ACIS-OBF, ACIS-CCD ACIS-CCD ACIS-OBF ACIS-CCD ACIS-CCD ACIS-CCD ACIS-CCD ACIS-CCD HRMA ACIS-CCD HRMA ACIS-CCD ACIS-CCD HRMA HRMA HRMA ACIS-CCD ACIS-CCD ACIS-CCD ACIS-CCD ACIS-CCD

a Additional instrumental spectral features arise due to X-ray absorption ïne structure produced in the CCDs and UV/optical blocking ïlter. These features extend for about a few hundred eV above each of absorption edges of Al-L, C-K, N-K, O-K, Al-K. b E is the energy of the incident photon and E \ 1.739 is the energy of the silicon 0 f ÿuorescence photons. c Mirror absorption energies from Graessle et al. 1993. d These energies are dependent on the gain and bias levels and will vary from chip to chip.

yields a best-ït temperature of 2.7 ^ 0.2 keV (LYNX ; ït 2 in Table 4) or 2.29`1.0 keV (XSPEC ; ït 4 in Table 4). ~0.5 Abundances were ïxed at 0.3 of the cosmic value. An F-test between ïts 1 and 2 indicates that neither model is signiïcantly preferred over the other. The 2õ10 keV X-ray luminosity of the inner jet region, assuming an absorbed power-law model (model 1 in Table 4), is 0.27 ] 1044 ergs s~1.

2.5. Spectral Analysis of the Outer Jet Component The X-ray spectra for the outer jet component were extracted from circles centered on knot WK8.9 with radii of 2A . Only standard ASCA grades 0, 2, 3, 4, and 6 were .5 included, and the background was determined by extracting events within annuli centered on the core with inner and outer radii of 15A and 25A, respectively. Pileup eects are

TABLE 4 MODEL PARAMETERS DETERMINED FROM SPECTRAL FITS TO THE CHANDRA ACIS-S SPECTRA OF THE INNER JET COMPONENT OF PKS 0637[752 N H (z \ 0) (1020 cm~2) 12.6`0.1 ~0.1 12.0`0.3 ~0.3 9.0 (ïxed) 9.0 (ïxed) Fluxb (10~15 ergs s~1 cm~2) 5.9 5.6 7.8 7.3 (8.6) (3.3) (6.8) (3.0) Lc X (1044 ergs s~1) 0.42 0.29 0.63 0.34 (0.29) (0.19) (0.25) (0.20)

Modela 1 2 3 4 ....... ....... ....... .......

! or T

e 2.0`0.2 ~0.2

T e (keV) 2.7`0.2 ~0.2

Flux Density at 1 keV (10~15 ergs s~1 cm~2 keV ~1) 4.1 3.8 5.1 4.7

2.29`1.0 ~0.5 a Fits 1 and 3 incorporate a power-law spectrum plus absorption due to cold material at solar abundances ïxed to the Galactic value. Spectral ïts 2 and 4 incorporate a Raymond-Smith thermal plasma model with the abundance set at 0.3 of the cosmic value. Fits 1 and 2 were performed in the observed energy range of 0.2õ 4 keV using LYNX. Fits 3 and 4 were performed in the observed energy range of 0.6õ 4 keV using XSPEC. b Fluxes calculated in the ranges 0.2õ2 keV and 2õ10 keV (quoted in parentheses). X-ray ÿuxes are not corrected for Galactic absorption. c Luminosities calculated in the ranges 0.2õ2 keV and 2õ10 keV (quoted in parentheses). Luminosities are in the rest frame and are corrected for Galactic absorption.

2.27`0.2 ~0.2

s2/(dof ) l 1.5 (14) 1.3 (14) 1.9 (8) 2.0 (8)


No. 2, 2000

CHANDRA RESOLVES JET IN PKS 0637[752

663

FIG. 7.õSpectra and best-ït models of the inner (ït 3 from Table 4) and outer (ït 3 from Table 5) jet components of PKS 0637[752

negligible for the outer jet component because of the extended nature of the emission combined with the relatively low count rate of 0.025 counts s~1. We combined a subset of the observations of PKS 0637[752 listed in Table 1 to produce spectra of total exposure 32,931 s for node 0 of S3 and 70,700 s for node 1 of S3. The XSPEC and LYNX spectral ïts were performed in the 0.6õ 6 keV range and 0.2õ 6 keV range, respectively. Plausible emission mechanisms for the production of the observed X-rays from the knots are synchrotron self-Compton and thermal bremsstrahlung emission from a compressed shocked medium. To distinguish between possible emission mechanisms, we ït the composite knot spectrum with absorbed power-law and thermal-plasma models. More complex models were not pursued because of the relatively low counts in the composite spectrum (see Table 1) and the present uncertainties in the low-energy instrumental response. Our results are pre-

sented in Table 5. Spectral ïts with thermal and power-law models provide similar reduced s2 . The 2õ10 keV lumil nosity of the outer jet region, assuming an absorbed powerlaw model (ït 1 in Table 5), is 2.2 ] 1044 ergs s~1. No signiïcant emission lines are detected in the outer jet spectrum. The composite spectrum for the outer jet of PKS 0637[752 with best-ït model (ït 3 in Table 5) is shown in Figure 7. 2.6. Properties of Sources in the Near V icinity of PKS 0637[752 Several relatively X-ray bright sources (count rates above 2 ] 10~3 counts s~1) were detected on CCD S3 in the vicinity of PKS 0637[752. In Table 6 we list their coordinates, count rates, and distances from the core of PKS 0637[752. We searched the USNO catalog and found optical counterparts within 1A in only ïve out of the 12 sources. No

TABLE 5 MODEL PARAMETERS DETERMINED FROM SPECTRAL FITS TO THE CHANDRA ACIS-S SPECTRA OF THE OUTER JET COMPONENT OF PKS 0637[752 N H (z \ 0) (1020 cm~2) 11.8`0.3 ~0.3 11.8`0.1 ~0.1 9.0 (ïxed) 9.0 (ïxed) Fluxb (10~13 ergs s~1 cm~2) 0.40 0.43 0.50 0.48 (0.75) (0.41) (0.82) (0.61) Lc X (1044 ergs s~1) 2.3 2.0 2.7 1.9 (2.3) (1.9) (2.5) (2.4)

Fita 1 2 3 4 ...... ...... ...... ......

! 1.83`0.1 ~0.1 1.85`0.08 ~0.08

T (keV) 3.8`0.2 ~0.2

Flux Density at 1 keV (10~14 ergs s~1 cm~2 keV~1) 2.8 3.0 3.3 3.2

5.6`1.1 ~0.6 a Fits 1 and 3 incorporate a power-law spectrum plus absorption due to cold material at solar abundances ïxed to the Galactic value. Fits 2 and 4 incorporate a Raymond-Smith thermal plasma model plus absorption due to cold material. Metal abundances were held ïxed at 0.3 for ïts 2 and 4. Spectral ïts 3 and 4 were performed with XSPEC in the observed energy range of 0.6õ7 keV with simultaneous ïts to spectra extracted from nodes 0 and 1. Fits 1 and 2 were performed with LYNX in the observed energy range 0.2õ7 keV with events extracted from node 1 only. b Fluxes calculated in the ranges 0.2õ2 keV and 2õ10 keV (quoted in parentheses). X-ray ÿuxes are not corrected for Galactic absorption. c Luminosities calculated in the ranges 0.2õ2 keV and 2õ10 keV (quoted in parentheses). Luminosities are in the rest frame and are corrected for Galactic absorption.

s2/(dof ) l 2.0 (49) 1.9 (49) 1.3 (85) 1.4 (85)


664

CHARTAS ET AL.
TABLE 6 SOURCES IN THE NEAR VICINITY OF PKS 0637[752 X (pixel) ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... 4149 3655 3705 3897 4667 4047 3019 4051 3925 3745 4275 4441 Y (pixel) 4209 4007 3987 3257 4053 4173 4185 3683 3729 4185 3483 3605 R.A. (J2000) 6 6 6 6 6 6 6 6 6 6 6 6 35 35 36 36 36 35 36 34 34 35 35 34 51.5 38.3 30.3 07.2 35.5 51.2 42.1 46.7 31.6 00.3 22.0 38.0 Decl. (J2000) [75 [75 [75 [75 [75 [75 [75 [75 [75 [75 [75 [75 15 15 15 19 16 19 16 15 16 20 21 19 28 10 22 06 59 29 49 17 27 06 07 21 Distancea (arcmin) 0.89 1.25 2.98 3.16 3.25 3.26 3.63 4.01 4.85 4.90 5.17 5.40 Count Rate (10~2 counts s~1) 0.71 1.19 0.37 0.27 0.56 0.26 1.24 0.30 0.64 0.29 0.36 0.23 ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ 0.06 0.08 0.04 0.04 0.05 0.04 0.08 0.04 0.06 0.04 0.04 0.03

Vol. 542

Object CXO CXO CXO CXO CXO CXO CXO CXO CXO CXO CXO CXO J063551.5[751528 J063538.3[751510 J063630.3[751522 J063607.2[751906 J063635.5[751659 J063551.2[751929 J063642.1[751649 J063446.7[751517 J063431.6[751627 J063500.3[752006 J063522.0[752107 J063438.0[751921

NOTE.õUnits of right ascension are hours, minutes, and seconds, and units of declination are degrees, arcminutes, and arcseconds. a Distance from core of PKS 0637[752.

counterparts were found in the NED and SIMBAD catalogs.
3.

DISCUSSION

In Figure 8, we present the SED of the WK7.8 knot of PKS 0637[752. The radio observations of PKS 0637[752 were performed at ATCA at 4.8 and 8.6 GHz. The 4.8 and 8.6 GHz beam width is D2A and D1A FWHM, respectively. The spectral indices and ÿux densities of the resolved components of the core, jet, and knots are presented in Table 7. Values for the optical ÿux density were obtained from the recent Hubble Space T elescope WFPC2 observations (Schwartz et al. 2000 and references therein).

The SED shows that a single-component power-law synchrotron model cannot explain the combined radio, optical, and X-ray ÿux densities since the optical lies far below a power-law interpolation between the radio and X-ray measurements. We also tested whether SSC emission or inverse Compton scattering of cosmic microwave background (CMB) photons could explain the observed X-ray emission. The model components in Figure 8 are for the case of equipartition between the magnetic ïeld and electron energy densities and assume a sphere of radius 0A 5 and a power.1 law electron number spectrum of slope 2.4 between 100 MeV and 230 GeV, steepening by unity at 30 GeV due to energy losses. A detailed description of the SSC and inverse Compton calculations and the assumptions made for the model parameters are presented in a companion paper (Schwartz et al. 2000). We estimate an equipartition ïeld, B , of about 2 ] 10~4 G. Based on this B -value, both eq SSC and IC on the CMB underpredict the eq X-ray ÿux by several orders of magnitude. Thermal models were also considered. Assuming a plasma temperature of 10 keV, an emission volume of 4 ] 10~3 arcsec3 anda2õ10 keV luminosity of 1 ] 1042 ergs s~1, we estimate a plasma density of about 1 cm~3. The derived rotation measures (RMs) for radio waves propagating through such a dense plasma are quite large, inconsistent with the recent radio ATCA observations at 4.8 and 8.6 GHz (Schwartz et al. 2000 ; Lovell et al. 2000). These observations show RM D 80 rad m~2 in the
TABLE 7 RADIO SPECTRAL INDICES AND FLUX DENSITIES OF PKS 0637[752 COMPONENTS 4.8 GHz Flux Density (Jy) 6.373 0.321 0.167 0.206 7.03 ^ 0.02 8.6 GHz Flux Density (Jy) 6.343 0.200 0.095 0.110 6.72 ^ 0.02

Component Core ..................... Inner west jet ........... Outer west jet .......... East jet .................. Total cleaned ÿux ......

E [0.01 0.81 0.97 1.08

a

FIG. 8.õSpectral energy distribution for knot WK7.8 in the PKS 0637[752 jet. The solid line is the synchrotron component, the dashed line is the SSC component, and the dotted line is the Compton-scattered CMB component, as discussed in ° 3 and Schwartz et al. (2000). The model components are based on equipartition assumptions.

NOTES.õAll radio spectral energy indices follow the convention of ÿux density Pl~aE , where a \ (! [ 1). Note that the inner and outer west E radio jets are deïned as the components of the jet before and after the bend at WK9.7 ; 1 kJy is equivalent to 10~29 ergs cm~2 s~1 Hz~1.


No. 2, 2000

CHANDRA RESOLVES JET IN PKS 0637[752

665

core but no Faraday rotation in the jet, with an upper limit of ^10 rad m~2. A contrived geometry where the jet collides with a giant molecular cloud in a companion galaxy producing thermal X-rays, but from our line of sight the cloud is located behind the jet, may explain the nondetection of Faraday rotation in the jet. An examination of the radio and X-ray brightnesses of dierent parts of the main emission region in the jet suggests that the X-ray brightness to radio brightness ratio is remarkably constant out to the last knot (WK9.7), where the X-rays are relatively fainter. The interpretation of the change of X-ray brightness between WK8.9 and WK9.7 depends, however, on the emission process : if the Xradiation has a synchrotron origin, then the emitting electrons must be locally accelerated, and the change in X-ray brightness of WK9.7 would be telling us about changes in particle acceleration at dierent points in the jet. If the X-rays have an inverse Compton origin, then it is possible that the brightness change is entirely due to aging of an electron population accelerated in WK8.9, but then other difficulties in understanding the energetics of the source must be faced (see Schwartz et al. 2000 for further discussion).
4.

(Schwartz et al. 2000 ; Lovell et al. 2000) have been reanalyzed to search for compact components close to the radio knots, and we ïnd that the knots are indeed resolved at 0A 5 resolution at 5 GHz, with less than 5 mJy remaining .0 at this resolution. This suggests that they are low surface brightness "" hot spots.îî The spectral analysis of the core, jet, and knot components has been quite complex due to the dierent nonoptimal conïgurations used for each observation and the present uncertainty in several of the Chandra/ACIS calibration data sets. Having quoted the above caveat, we summarize the spectral analysis of the core and jet as follows : 1. The core ÿux and spectral shape are consistent with those measured with ASCA. However, we do not detect the emission line near 1 keV claimed in a recent ASCA observation of PKS 0637[752. The HPR for ASCA is about 3@, so one possible explanation for this discrepancy is that the emission line claimed to be detected with ASCA may originate from a nearby unresolved source. We have detected three relatively weak X-ray sources within 3@ of PKS 0637[752. However, the spectra of these sources do not show any prominent 1 keV emission lines that could explain the ASCA results. 2. The X-ray spectrum of the inner jet component appears to be slightly steeper than that of the outer jet region (see Tables 4 and 5). A dierence in spectral slopes may be explained as follows : A population of synchrotronemitting relativistic electrons in the inner jet region are undergoing radiation losses and producing the observed steeper spectra (X-ray and radio spectral indices a D E [1.0). As these electrons enter the outer jet region (6A õ11A away from core), they are reaccelerated in a shock .5 .5 or some other structure to produce a ÿatter spectrum (X-ray and radio spectral indices a D [0.8). As they age past E 11A , the radio spectrum steepens again (radio spectral .5 .5 index a D [1.0). Also, in the region 11A away from the E core where the highest energy electrons may plausibly have lost all their energy, the X-rays turn o as expected. 3. Spectral ïts to the outer jet spectrum assuming thermal and power-law models yield similar s2 values. The outer jet region is relatively bright in X-rays with a 2õ10 keV luminosity of 2.2 ] 1044 ergs s~1 (ït 1 in Table 5). The best-ït value of the X-ray luminosity of the outer jet is not sensitive to the assumed emission process. The good agreement in X-ray and radio spectral slopes in the inner and outer jet regions strongly suggests that there is substantial electron acceleration in the knot complexes WK7.8 and WK8.9.

CONCLUSIONS

Chandraîs unique resolving-power capabilities open a new era in X-ray astronomy. We anticipate that many more radio jets and knots will be resolved in future observations with Chandra. The simultaneous spatial and spectral information provided with the Chandra/ACIS combination allows for accurate estimates of the size and spectral densities of the knots in extragalactic radio jets, which leads to tighter constraints on models that attempt to explain the X-ray emission. The standard emission processes usually invoked to explain X-ray emission from jets cannot explain the X-ray observations of PKS 0637[752. In particular, simple synchrotron models and equipartition SSC models underpredict the X-ray ÿux of the knots by many orders of magnitude. Thermal models predict shocked plasma densities and rotation measures that are too large. Particular contrived geometries of the jet and interacting molecular cloud may, however, explain the nondetection of Faraday rotation in the radio observations. More complex models that invoke inhomogeneities and/or nonequipartition and/or an extra photon source to explain the X-ray emission as inverse Compton are presented in Schwartz et al. (2000). A careful spatial analysis combining most of the available observations of PKS 0637[752 has resolved the jet and at least four knots along the jet. The X-ray knots are 5A , 7A , .7 .8 8A , and 9A from the core of PKS 0637[752, in good .9 .7 agreement with the locations seen in the radio image. The X-ray knots are not individually resolved, and the upper limit on their diameter is D0A . The radio knots from the .4 ATCA image are unresolved with an upper limit on their .4 diameter also of D0A . However, our VLBI observations

We would like to thank Eric Feigelson for helpful comments, Martin Hardcastle for software used to generate Figure 8, and Kenneth Lanzetta for providing the optical ÿux density of knot WK7.8 used in Figure 8. This work was supported by NASA grant NAS 8-38252.

APPENDIX LYNX : A SIMULATED-BASED SPECTRAL FITTING TOOL Astrophysical X-ray spectra are commonly analyzed by creating parameterized models for the incident spectra, folding these models through the telescope and instrument responses, and then adjusting the parameters by minimizing a metric such


666

CHARTAS ET AL.

as s2 formed between the observed and modeled spectra. A tool that uses this approach and is widely used to ït astrophysical X-ray spectra observed from a variety of X-ray satellites is XSPEC. The telescopeîs eective area dependence with o-axis angle combined with the detectorîs quantum efficiency dependence with energy are usually incorporated into an auxiliary response ïle while the response of a detector to monoenergetic photons of energy E are incorporated into a spectral redistribution matrix. The spectral redistribution matrix is usually created from parameterizing the output of CCD simulations of input monoenergetic spectra. One approximation of forward ïtting CCD spectra using telescope and detector response matrices is that the simulated CCD spectra are uniquely deïned for a given model and set of input model parameters. In reality, however, several physical processes within CCDs are nondeterministic such as ÿuorescent yields, absorption depths, and photon escape probabilities, and detected spectra for identical incident spectra will in general be slightly dierent. This eect becomes more noticeable for spectra containing a low number of counts. We have developed the tool LYNX that employs the forward ïtting approach to infer incident astrophysical spectra. However, it diers from the conventional deterministic tools, such as XSPEC, in that the mirror and detector characteristics are determined by incorporating Monte Carlo simulators. In particular, LYNX links to the ray-trace tool MARX (Wise et al. 1997) to simulate the mirror response and to the PSU ACIS simulator (Townsley et al. 2000, in preparation) to provide the CCD response. Astrophysical spectra obtained with ACIS are initially ït with the standard X-ray spectral ïtting package XSPEC to provide an initial guess for LYNX. The modeled incident spectrum is propagated through the Chandra mirrors and ACIS components with the MARX and the PSU ACIS simulators, respectively. A merit function that incorporates the dierences between the observed and simulated spectra spectrum is minimized using a downhill simplex method to yield the best-ït model parameters. LYNX simulates the propagation of individual photons through the HRMA/ACIS conïguration and also takes into account the possible overlap of the resulting charge clouds within each exposure. The spectra produced through LYNX therefore will simulate pileup. The present version of LYNX also allows ïtting spectra of any grade selection, corrects for vignetting, and accounts for the dither motion of the source across the CCD.
REFERENCES Arnaud, K. A. 1996, in ASP Conf. Ser. 101, Astronomical Data Analysis Meisenheimer, K., Roser, H.-J., Hiltner, P. R., Yates, M. G., Longair, M. S., Software and Systems V, ed. G. H. Jacoby & J. Barnes (San Francisco : Chini, R., & Perley, R. A. 1989, A&A, 219, 63 ASP), 17 Schwartz, D. A., et al. 2000, ApJ, 540, L69 Biretta, J. A., Stern, C. P., & Harris, D. E. 1991, AJ, 101, 1632 van Speybroeck, L. P., Jerius, D., Edgar, R. J., Gaetz, T. J., Zhao, P., & Dickey, J. M., & Lockman, F. J. 1990, ARA&A, 28, 215 Reid, P. B. 1997, Proc. SPIE, 3113, 89 Elvis, M., & Fabbiano, G. 1984, ApJ, 280, 91 Weisskopf, M. C., & OîDell, S. L. 1997, Proc. SPIE, 3113, 2 Graessle, D. E., et al. 1993, Proc. SPIE, 1742, 203 Wise, M. W., Davis, J. E., Huenemoerder, Houck, J. C., Dewey, D., FlanaHarris, D. E., Carilli, C. L., & Perley, R. A. 1994, Nature, 367, 713 gan, K. A., & Baluta, C. 1997, The MARX 2.0 User Guide, CXC Internal Harris, D. E., et al. 2000, ApJ, 530, L81 Document Harris, D. E., & Stern, C. P. 1987, ApJ, 313, 136 Yaqoob, T., George, I. M., Turner, T. J., Nandra, K., Ptak, A., & SerleLawson, A. J., Turner, M. J. L., Williams, O. R., Stewart, G. C., & Saxton, mitsos, P. J. 1998, ApJ, 505, L87 R. D. 1992, MNRAS, 259, 743 Lawson, A. J., & Turner, M. J. L. 1997, MNRAS, 288, 920 Lovell, J. E. J., et al. 2000, in Astrophysical Phenomena Revealed by Space VLBI, ed. H. Hirabayashi, P. G. Edwards, & D. W. Murphy (Sagamihara : ISAS), 215