1: \documentclass[manuscript]{aastex}

2: %\documentclass[manuscript]{/local/data/kea1/rms/aastex/aastex52/aastex}

3: %\documentstyle[aasms4]{article}

4: %\documentclass[12pt,preprint]{aastex}

5: %\documentclass{aastex}

6: %\usepackage{emulateapj5}

7: 

8: \def\oiii{[{\sc O\, iii}]}

9: \def\oii{[{\sc O\, ii}]}

10: \def\mgii{[{\sc Mg\, ii}]}

11: \def\feka{Fe K$\alpha$}

12: \def\chandra{{\it Chandra}}

13: \def\xmm{{\it XMM-Newton}}

14: \def\asca{{\it ASCA}}

15: \def\hst{{\it HST}}

16: \def\rxte{{\it RXTE}}

17: \def\sax{{\it BeppoSAX}}

18: \def\ginga{{\it Ginga}}

19: \def\rosat{{\it ROSAT}}

20: \def\ein{{\it Einstein}}

21: \def\integral{{\it INTEGRAL}}

22: \def\merlin{{\it MERLIN}}

23: \def\vla{{\it VLA}}

24: \def\vlba{{\it VLBA}}

25: \def\glast{{\it GLAST}}

26: \def\swift{{\it Swift}}

27: \def\suzaku{{\it Suzaku}}

28: \def\sirtf{{\it Spitzer}}

29: \def\lum{erg s$^{-1}$}

30: \def\flux{erg cm$^{-2}$ s$^{-1}$}

31: \def\nh{cm$^{-2}$}

32: \def\arcsec{$^{\prime\prime}$}

33: \def\deg{$^{\circ}$}

34: 

35: \def\ltsima{$\; \buildrel < \over \sim \;$}

36: \def\simlt{\lower.5ex\hbox{\ltsima}} % < over ~

37: \def\gtsima{$\; \buildrel > \over \sim \;$}

38: \def\simgt{\lower.5ex\hbox{\gtsima}} % > over ~

39: 

40: \def\cl{\centerline}

41:  

42: %\input{/local/data/kea1/rms/aastex/psfig.sty}

43: \input{psfig.tex}

44: 

45: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

46: %

47: %   FIGURES:   1: BAT significance image 

48: %              2: SDSS and HET spectrum 

49: %              3a-b: BAT and XRT average light curve

50: %              4a-d: the four XRT light curves 

51: %              5: the XRT residuals 

52: %              6: SED 

53: %              7: SED in context 

54: %

55: %   TABLES:    1- log of Swift observations 

56: %              2- Spectral results 

57: %              3- The optical line props 

58: %              4- spectral fitting results 

59: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

60: 

61: 

62: \begin{document}

63: \title{Discovery of an Extreme MeV Blazar with the Swift Burst Alert Telescope} 

64: 

65: %\normalsize

66:  

67: 

68: \author{R.M. Sambruna, C. B. Markwardt, R.F. Mushotzky, J. Tueller, R.

69: Hartman}

70: \affil{NASA's GSFC, Greenbelt, MD 20771}

71: 

72: \author{W.N. Brandt, D.P. Schneider, A. Falcone, A. Cucchiara} 

73: \affil{Department of Astronomy \& Astrophysics, The Pennsylvania State

74: University, University Park, PA 16802} 

75: 

76: \author{M. F. Aller, H. D. Aller} 

77: \affil{University of Michigan, Ann Arbor, MI 48109-1042} 

78: 

79: \author{I. Torniainen}

80: \affil{ Mets\"ahovi Radio

81: Observatory,  Helsinki University of Technology, Mets\"ahovintie 114,

82: FIN-02540 Kylm\"al\"a, Finland}

83: 

84: \author{F. Tavecchio, L. Maraschi}

85: \affil{Osservatorio Astronomico di Brera, via Brera 28, 20121 Milano,

86: Italy}

87: 

88: \author{M. Gliozzi}

89: \affil{George Mason University, Dept. of Physics \& Astronomy and School of

90: Computational Sciences, MS 3F3, 4400 University Drive, Fairfax, VA 22030}

91: 

92: \author{T. Takahashi} 

93: \affil{Institute of Space and Astronautical Science, 3-1-1 Yoshinodai,

94: Sagamihara-shi, Kanagawa 229-8510, Japan}

95: 

96: 

97: \begin{abstract}

98: 

99: The Burst Alert Telescope (BAT) onboard \swift\ detected bright

100: emission from 15--195~keV from the source SWIFT~J0746.3+2548 (J0746 in

101: the following), identified with the optically-faint (R $\sim$ 19),

102: $z$=2.979 quasar SDSS J074625.87+244901.2. Here we present \swift\ and

103: multiwavelength observations of this source. The X-ray emission from

104: J0746 is variable on timescales of hours to weeks in 0.5--8~keV and of

105: a few months in 15--195~keV, but there is no accompanying spectral

106: variability in the 0.5--8 keV band. There is a suggestion that the BAT

107: spectrum, initially very hard (photon index $\Gamma \sim 0.7$),

108: steepened to $\Gamma \sim 1.3$ in a few months, together with a

109: decrease of the 15--195~keV flux by a factor $\sim$ 2. The 0.5--8~keV

110: continuum is well described by a power law with $\Gamma \sim 1.3$, and

111: spectral flattening below 1~keV. The latter can be described with a

112: column density in excess of the Galactic value with intrinsic column

113: density N$_H^z \sim 10^{22}$ \nh, or with a flatter power law,

114: implying a sharp ($\Delta \Gamma$ \gtsima 1) break across 16~keV in

115: the quasar's rest-frame. The Spectral Energy Distribution of J0746 is

116: double-humped, with the first component peaking at IR wavelengths and

117: the second component at MeV energies. These properties suggest that

118: J0746 is a a blazar with high gamma-ray luminosity and low peak energy

119: (MeV) stretching the blazar sequence to an extreme. 

120: 

121: \end{abstract}

122: 

123: \keywords{Galaxies: active --- galaxies: 

124: quasars: individual (J0746.3+2548)--- X-rays: galaxies} 

125: 

126: \section{Introduction}

127: 

128: Among Active Galactic Nuclei (AGN), blazars are characterized by the

129: most extreme properties, such as rapid variability, large apparent

130: isotropic luminosity, high and variable polarization in the radio and

131: optical, and emission across the entire electromagnetic spectrum, from

132: radio wavelengths to GeV and TeV  gamma-ray energies (e.g., Ulrich,

133: Maraschi, \& Urry 1997). These properties are ascribed to non-thermal

134: (synchrotron and inverse Compton) emission from a relativistic jet 

135: seen nearly end-on. Blazars' Spectral Energy Distributions (SEDs) are

136: double-humped in a power-per-decade representation, with the

137: synchrotron component peaking anywhere from IR to X-rays, and the

138: inverse Compton component extending up to gamma-rays. The SEDs form a

139: sequence in luminosity, with more luminous sources having both peaks

140: at lower energies than their fainter counterparts (Fossati et

141: al. 1998, Ghisellini et al. 1998). 

142: 

143: Traditionally, blazars have been selected at radio, optical, and X-ray

144: wavelengths according to various criteria (e.g., Laurent-Mueheleisen

145: et al. 1997, Perlman et al. 1998). Only a few sources were discovered

146: using gamma-rays, both at GeV (Romani et al. 2004) and MeV wavelengths

147: (Blom et al. 1995). The advent of detectors with improving sensitivity

148: at hard X-ray energies ($>$ 10 keV), is now providing a new window for

149: blazar discovery and observation.

150:  

151: Such a window was recently opened by the Burst Alert Telescope (BAT;

152: Barthelmy et al. 2005) onboard the \swift\ satellite. On December 15,

153: 2004, the BAT began an ongoing monitoring of the sky with its

154: wide-angle (2 steradians) field of view and high sensitivity in

155: the 15--200~keV energy band. During the first 3 months of the sky survey

156: operations, with a flux threshold of $\sim 10^{-11}$ \flux, the BAT

157: detected several AGN, both radio-loud and radio-quiet (Markwardt et

158: al. 2006). Here we describe the BAT discovery and subsequent \swift\

159: and multiwavelength observations of a new blazar with extreme

160: properties, SWIFT~J0746.3+2548 (J0746 in the following). The source

161: was identified with an optically-faint quasar at $z$=2.979 in the

162: Sloan Digital Sky Survey (SDSS) archive. In this paper, we will

163: present evidence that J0746 is a blazar (Flat Spectrum Radio Quasar)

164: and more specifically, a new member of the MeV blazar

165: class\footnote[1]{We define an MeV blazar as a blazar with peak of the

166: Compton component in the SED at MeV energies.}. 

167: 

168: In Fall 2005, J0746 was the target of an X-ray (\swift\ and

169: \suzaku), optical, and radio campaign. 

170: The radio observations were taken with the 26-m diameter U. of

171: Michigan Radio Astronomy Observatory (UMRAO) and the 14-m diameter

172: Mets\"ahovi radio telescopes, while optical-UV photometry is from the

173: \swift\ UVOT telescope.  Here we present the \swift\ and radio

174: observations, while a complete account of the

175: \suzaku\ data will be forthcoming (Takahashi et al. 2006, in prep.). We also

176: report on a new optical spectrum taken with the {\it Hobby-Eberly Telescope} (HET).

177: 

178: This paper is structured as follows. In \S~2 we describe the source

179: discovery and identification. The observations and data reduction

180: procedures are presented in \S~3, while in \S~4 we give the results,

181: in \S~5 the discussion, and in \S~6 the conclusions. We assume a

182: concordance cosmology with H$_0=71$ km s$^{-1}$ Mpc$^{-1}$,

183: $\Omega_{\Lambda}$=0.73, and $\Omega_m$=0.27 (Spergel et

184: al. 2003). The energy index $\alpha$ is defined such that $F_{\nu}

185: \propto \nu^{-\alpha}$. 

186:  

187: \section{Discovery of J0746 at hard X-rays} 

188: 

189: During the first 3 months of its sky survey, the BAT discovered hard

190: X-ray emission from several extragalactic sources (Markwardt et

191: al. 2006). With a flux F$_{20-200~keV} \sim 9 \times 10^{-11}$ \flux\

192: and a $>7\sigma$ detection (Fig. 1), SWIFT~J0746.3+2548 stands out as

193: one of the brightest. The follow-up observations with the XRT in

194: 0.5--10 keV shows the presence of a source at RA(2000)= 07hrs 46m

195: 25.6s, DEC(2000)=+25deg 49m 02.0s, or 1\arcmin\ from the BAT

196: position. The radii of the error circles for the BAT and XRT positions

197: are 2\arcmin\ and 3\arcsec, respectively. No other X-ray sources are

198: present in the XRT field of view, leading us to identify the XRT

199: source with the origin of the BAT flux.

200: 

201: A comparison with the latest Sloan Digital Sky Survey (SDSS; York et

202: al. 2000) Quasar Catalog (Schneider et al. 2005) yields a match with

203: SDSS~J074625.87+254902.2, located at RA(2000)=07h46m25.9,

204: DEC(2000)=+25d49m02.1s; the accuracy of the SDSS coordinates is

205: approximately 0.1\arcsec.  The SDSS counterpart of the X-ray source is

206: an optically faint (R $\sim$ 19 mag) quasar at a redshift

207: $z$=2.979. Using optical (Richards et al. 2006) and X-ray (Bauer et

208: al. 2003) number counts, we estimate that the false-match probability

209: of the XRT and SDSS sources is $P\sim 6.2

210: \times 10^{-6}$. Thus, the identification of the \swift\ source with

211: SDSS~J074625.87+254902.2 (J0746 in the following) is regarded as

212: robust.

213: 

214: Because of its optical faintness, J0746 largely escaped the attention

215: of the Astronomy community. A search of the \verb+NED+ database shows

216: 23 references. Radio \vlba\ observations at 2.3 and 8.6 GHz show the

217: quasar to be core-dominated, with no obvious extended radio features

218: on mas scales (Fey \& Charlot 2003). From the data in this paper, we

219: calculate a radio-loudness parameter, $\log R_L

220: \sim 3$, in the rest-frame of the quasar. The radio-loudness parameter

221: is defined as the ratio of the 5~GHz to the $V$--band flux. To

222: K-correct the radio and optical fluxes we used power law continua with

223: $\alpha_{\nu, radio}=0.08$ (see \S~3.4) and $\alpha_{\nu,opt}=0.54$

224: (\S~3.3).

225: 

226: Although J0746 was only recently identified as a quasar, the archival

227: literature contains a number of observations of the object obtained

228: during wide-angle surveys at a number of wavelengths.  In the radio,

229: J0746 was monitored at 8 and 2.3~GHz with the Green Bank

230: Interferometer in the time period 1988--1995 (Lazio et al. 2001). No

231: variability was detected at 2~GHz. However, the light curve at 8~GHz

232: (32~GHz in the rest-frame) shows a slow but smooth increase of the

233: flux over 3 years, with a max/min flux change of a factor of 3. The

234: radio spectral index steepens on the same timescale (Fig. 8 of Lazio

235: et al. 2001).

236: 

237: The only previous X-ray observation of J0746 was by the \rosat\

238: satellite during its All-Sky Survey (RASS). Weak 0.2--2~keV emission,

239: with 16 counts in 438 sec, was recorded. Since the Galactic column

240: density in the direction of J0746 is low, N$_H^{Gal}=4.04

241: \times 10^{20}$ \nh, this implies a highly absorbed AGN;

242: alternatively, \rosat\ caught the quasar in a low intensity state. 

243: 

244: Based on its broad-band properties, we argue in this paper in favor of

245: a blazar nature of J0746 (\S~5). 

246: 

247: \section{Observations} 

248: 

249: \subsection{X-rays} 

250: 

251: The \swift\ observations of J0746 are summarized in Table~1, where the

252: BAT observations and follow-up XRT pointings are reported. The

253: November 4 XRT observation was performed contemporaneously with

254: \suzaku\ and radio measurements. 

255: 

256: \noindent{\bf BAT:} J0746 was in the

257: BAT field of view for a total of 755~ks since the beginning of the

258: survey on December 15, 2004. The results described here are for the

259: first 9 months of the survey, or until the end of September, 2005. The

260: average count rate from 15--195~keV is listed in Table~1.

261: 

262: The reduction of the BAT data was described in Marqwardt et

263: al. (2005).  Figure~1 shows the BAT significance map, ranging from 

264: $-2\sigma$ (white) to $+8\sigma$ (black). Pixel size is

265: 5\arcmin. The cross marks the SDSS position of J0746. The source is

266: detected at $>7\sigma$.

267: 

268: We extracted two BAT spectra, one for the initial 3.5 months of the

269: survey, when J0746 was first discovered, and one integrated over the

270: whole 9 months. The BAT spectrum of J0746 is based on the mosaicked

271: sum of all observations which contain the source (mid Dec 2004 - mid

272: Sep 2005).  The minimum partial coding threshold was 15\%. Before

273: mosaicking, the images were corrected for systematic off-axis

274: attenuation effects, such that the spectrum should remain relatively

275: constant with off-axis angle (calibrated by BAT observations of the

276: Crab).

277: 

278: The BAT spectra were fitted with \verb+XSPEC+ v.11.3.2, with the

279: response matrix based on the latest (December 2005) calibration

280: products. Because of the spectral extraction protocol, the BAT

281: spectrum consists of 4 bins containing the source count rates in 4

282: energy bands: 14--25~keV, 25--50~keV, 50--100~keV, and 100--200~keV.

283: 

284: \noindent{\bf XRT:}

285: The XRT observed J0746 four times (Table~1). During the first three

286: observations in September, which overlapped with the BAT observations

287: (Fig.~2a), short exploratory exposures were used. On November 4, XRT

288: observed J0746 for 15~ks, together with the UVOT,

289: \suzaku, and ground-based radio observations. The XRT count rates in the

290: four epochs are listed in Table~1.

291: 

292: We used the event2 files provided by the \verb+HEASARC+, cleaned

293: according to standard criteria (Capalbi et al. 2005). Only data

294: acquired through Photon Counting (PC) mode, which has the maximum

295: sensitivity, were used; no pileup is present at the source's count

296: rates (Table~1). Light curves and spectra were extracted from a

297: circular region centered on the XRT position and with radius

298: 20\arcsec. The background was extracted from a circle with radius

299: 130\arcsec\ in a region free from obvious X-ray sources. Inspection of

300: the background light curve shows no variability within the four XRT

301: exposures.

302: 

303: The XRT spectra were rebinned in order to have at least 20 counts in

304: the new bins, except for the first observation on 2005 September 5

305: (OBSID 002) where the spectrum was rebinned with a minimum of 10

306: counts/bin, to increase the number of data points at lower energies

307: ($<1$ keV) for a meaningful determination of the column density. The

308: spectra were fitted within \verb+XSPEC+ v.11.3.2 with the calibration

309: files provided by the \verb+HEASARC+, in the energy range 0.5--8 keV,

310: where the calibration is best known and the background negligible.

311: 

312: \subsection{Optical/UV photometry} 

313: 

314: \noindent{\bf UVOT:} 

315: The \swift\ UVOT observed the source continuously (except for Earth

316: occultations) starting at UT Nov 4 00:39 until UT Nov 6 15:59

317: (Table~2).  During each orbit data were collected from the six filters,

318: from the Optical to Ultraviolet wavelengths.  Since the source is

319: faint, we coadded all the data collected on the first day (Nov 4) in

320: only one image for each filter. Similarly, the Nov 5 and Nov 6 data

321: were coadded. The log of the UVOT observations, together with the

322: fluxes, is given in Table~2.

323: 

324: The data analysis was performed using the \verb+uvotsource+ task

325: included in the latest HEASOFT software release

326: (http://swift.gsfc.nasa.gov/docs/software/lheasoft/).  The background

327: was subtracted, and corrections for coincidence loss effects (similar

328: to pileup for the XRT) were applied. The magnitudes are converted into

329: fluxes using the latest in-flight flux calibration factors and

330: zero points.  The source is well detected in the optical filters with

331: over $3\sigma$ confidence level.  There is no detection in the UV

332: images, and only upper limits are reported in Table~2.

333: 

334: \noindent{\bf SDSS:} Table~2 presents optical photometric 

335: measurements of J0746 obtained from the SDSS DR3 catalog in the SDSS

336: system (Fukugita et al. 1996) on 19 December 2001. All fluxes are

337: dereddened for Galactic absorption. The SDSS fluxes are consistent

338: with those from the UVOT at similar wavelengths.

339: 

340: A search of the USNO archive for J0746 returns the following

341: historical records of its optical flux: $B$=20.01, 19.87 mag;

342: $R$=19.16, 19.04 mag; and I=18.92 mag. The $B$-band measurements are

343: consistent with those from the UVOT. 

344: 

345: \subsection{Optical spectra} 

346: 

347: \noindent{\bf SDSS:} 

348: The SDSS spectrum of J0746 is displayed in the top panel of

349: Figure~2. The data have a spectral resolution R $\sim$ 1900. 

350: 

351: The optical spectrum of J0746 is that of a typical $z \approx 3$

352: quasar. Broad prominent emission lines, such as Ly$\alpha$ (4827 \AA),

353: C~IV 1549~\AA\ (6164~\AA), and Si~IV/O~IV 1400~\AA\ (5571 \AA),

354: typical of high-$z$ quasars, are apparent in the spectrum. The

355: continuum power law slope in the SDSS spectrum is $\alpha{\nu} \sim

356: 0.54$, very near the canonical $\alpha{\nu}=0.5$ from the SDSS quasar

357: composite (Vanden Berk et al. 2001).

358: 

359: We calculated the observed EWs and fluxes of the detected lines. For

360: C~IV we find EW$_{obs}$=31 \AA\ and F$_{\lambda} \sim 9.8 \times

361: 10^{-16}$ \flux \AA$^{-1}$, while for Si~IV/O~VI EW$_{obs}$=78 \AA\

362: and F$_{\lambda} \sim 2.5 \times 10^{-15}$ \flux \AA$^{-1}$,

363: respectively. The EWs are within the distribution for radio-loud and

364: radio-quiet quasars (Wills et al. 1993). There is no evidence for C~IV

365: absorption, either in the emission line or in the neighboring

366: continuum.

367: 

368: \noindent{\bf HET:} Blazars have notoriously variable levels of

369: optical continuum which may dilute the emission lines during high

370: states. We thus acquired a second optical spectrum of J0746 in 2005

371: Fall with the Hobby-Eberly Telescope (Ramsey et al. 1994). The HET

372: spectrum of J0746 was obtained on 10 October 2005 with the Low

373: Resolution Spectrograph (Hill et al. 1998).  The 1.5\arcsec\ slit, 300

374: line mm$^{-1}$ grating, and GG385 filter produced a spectrum from

375: 4000-8000~\AA\ at a resolution R=460.  The spectrum, which is the sum

376: of two 900~s exposures, is shown in the lower panel of Figure~2.

377: 

378: The HET spectrum, taken 3.7 years (only 11 rest-frame months) after

379: the SDSS one, shows very similar features. The HET continuum is redder

380: and the lines slightly stronger, but these changes should be

381: considered marginal at best; the quality of the spectrophotometry of

382: the SDSS data is considerably higher than that of the HET observation,

383: and the change in the equivalent widths is not much more than

384: 1$\sigma$. Thus, we can conclude that the level of the optical

385: continuum did not significantly change between the epochs of the SDSS

386: and HET observations.

387: 

388: \subsection{Radio} 

389: 

390: The radio observations are presented in Table~3. 

391: 

392: \noindent{\bf UMRAO:} 

393: The University of Michigan radio data were obtained using the UMRAO

394: 26-meter prime focus paraboloid antenna which is equipped with

395: wide-band room temperature HEMPT (High Electron Mobility Pseudomorphic

396: Transistor) amplifiers and operates at central frequencies of 4.8,

397: 8.0, and 14.5~GHz. Measurements at all three of these frequencies

398: utilize rotating, dual-horn polarimeter feed systems. An on-off

399: observing technique was used at 4.8~GHz, and an on-on technique

400: (switching the target source between the two feed horns closely spaced

401: on the sky) was employed at 8 and 14.5~GHz. A typical observation

402: consisted of a series of 8 to 16 individual measurements over a 25 to

403: 45 minute period, depending on the observing frequency.

404: 

405: Observations of the program source were preceded and followed by

406: observations of nearby calibrator sources (3C 144, 3C 145, 3C 218, 3C

407: 274) selected from a grid across the sky; these calibration

408: measurements were used to correct for temporal changes in the antenna

409: aperture efficiency.  Frequent drift scans are also made across

410: stronger sources throughout every run to verify the telescope pointing

411: correction curves. The flux scale for the UMRAO observations is set by

412: observations of Cassiopeia~A (e.g., see Baars et al. 1977).  A more

413: detailed explanation of the calibration and analysis techniques is

414: given in Aller et al. (1985).  Interpolation of the radio fluxes in

415: Table~3 yields a slope $\alpha_r=0.08$, typical of a core-dominated

416: source as shown by the \vlba\ images at 2.3 GHz and 8.5 GHz (Fey \&

417: Charlot 2003). Comparing our 8~GHz measurements with the long-term

418: 8~GHz light curve in Fig. 8 of Lazio et al. (2001) shows that at the

419: time of the \swift\ observations in November 2005 the source was in a

420: high intensity state in the radio, similar to the maximum recorded in

421: late 1992. 

422: 

423: \noindent{\bf Mets\"ahovi:} 

424: Observations were also made with the 14-m diameter Mets\"ahovi radio

425: telescope at the frequency of 37 GHz. The fluxes are listed in

426: Table~3. 

427: 

428: The 37 GHz observations were made with the 13.7 m diameter Mets\"ahovi

429: radio telescope, which is a radome enclosed paraboloid antenna

430: situated in Finland (24 23' 38''E, +60 13' 05''). The measurements

431: were with a 1 GHz-band dual-beam receiver centered at 36.8 GHz. The

432: HEMPT front end operates at room temperature. The observations are

433: ON--ON observations, alternating the source and the sky in each feed

434: horn. A typical integration time to obtain one flux density data point

435: is 1200--1400 s.

436: 

437: The flux density scale is set by observations of DR 21. Sources 3C 84

438: and 3C 274 are used as secondary calibrators. A detailed description

439: of the data reduction and analysis is given in Ter\"asranta et

440: al. (1998). The error estimate in the flux density includes the

441: contribution from the measurement rms and the uncertainty of the

442: absolute calibration.

443: 

444: \section{X-ray results} 

445: 

446: \subsection{Timing analysis} 

447: 

448: Here we discuss the results of a timing analysis of the XRT and BAT

449: data.  Variability was assessed through the use of a $\chi^2$ test,

450: with related probability P$_{\chi^2}$ that the light curve is

451: constant; thus, small values of P$_{\chi^2}$ indicate significant

452: variability.

453: 

454: \noindent{\bf BAT:} 

455: The BAT light curve in the energy range 15--195~keV is shown in

456: Figure~3a. Bins are 15-day averages. The vertical ticks mark the dates

457: of the three September XRT observations. The BAT light curve is

458: formally consistent with a constant, with P$_{\chi^2}$=0.12.

459: 

460: \noindent{\bf XRT:}

461: The XRT light curves were extracted in soft (0.5--2~keV), hard

462: (2--8~keV), and total (0.5--8~keV) energy bands. We searched for

463: spectral changes by computing hardness ratios defined as the ratio of

464: the hard and soft X-ray light curves. 

465: 

466: Figure~3b shows the XRT long-term light curve from 0.5--8~keV,

467: obtained from the average count rates at the four observing epochs in

468: Table~1. From Figure~3b it is apparent that the flux changed by a

469: factor \gtsima 2 on timescales of a week. The $\chi^2$ test confirms

470: variability, with P$_{\chi^2} < 10^{-5}$. No spectral changes were

471: observed, with the plot of the hardness ratios versus time being

472: constant (P$_{\chi^2} \sim 0.098$). 

473: 

474: Next, we searched for variability within the XRT individual

475: exposures. Figure~4 shows the 0.5--8~keV light curves at the four

476: epochs, rebinned at 2000~sec resolution. Flux variations are observed

477: at every epoch, and most notably during the Sept 5 and Nov 4

478: observations. A $\chi^2$ test for these two light curves yields

479: significant variability, with P$_{\chi^2} < 10^{-3}$ and P$_{\chi^2}

480: \sim 9 \times 10^{-5}$, respectively. At the beginning of the Sept 5

481: observation the flux increases by a factor 2.5 in one hour, albeit

482: within large uncertainties. A similar increase of the flux on a

483: timescale of $\sim$ 2 hrs is also present at the beginning of the Nov

484: 4 observation. On this date, more modest variations are also seen

485: throughout the exposure, on similar timescales (1--2 ks).  No

486: significant spectral variability is detected in either the Sept 5 or

487: Nov 4 data, as shown by the the hardness ratio light curve

488: (P$_{\chi^2} > 0.06$).

489: 

490: We also investigated the energy-dependence of the variability by

491: calculating the variability amplitude relative to the mean count rate,

492: F$_{var}$, corrected for effects of random errors (e.g., Edelson et

493: al. 2002), for both hard and soft Nov 4 light curves. In practice,

494: however, due the paucity of data points and the large statistical

495: errors, only in few cases were we able to estimate F$_{var}$.  There

496: is a suggestion of a larger variability amplitude at harder X-rays,

497: with F$_{var, hard}$=0.160 and F$_{var, soft}$=0.05 for the hard,

498: although higher quality data are necessary to reach firmer conclusions.

499: 

500: \noindent{\bf Comparison to previous X-ray observations:} As mentioned

501: above, J0746 was detected during the RASS with a 0.2--2~keV count rate

502: of 0.04 c/s. To compare the \rosat\ and \swift\ observations, we

503: converted the RASS 0.2--2~keV count rate into an XRT 0.5--8~keV count

504: rate. Using \verb+PIMMS+ and assuming an absorbed power law with

505: column density N$_H=3 \times 10^{21}$ \nh\ and photon index

506: $\Gamma=1.3$ (see below), we find that the 0.5--8 keV count rate at

507: the time of the RASS observation was 0.041 c/s, comparable to the

508: lower state of the XRT observation in September 3 (Table~1). Thus, at

509: the time of the \rosat\ sky survey J0746 was in a lower X-ray

510: intensity state.

511: 

512: We conclude that the medium-hard X-ray emission of J0746 is

513: variable. There is a hint for larger-amplitude variations at the

514: harder energies, which will be tested in the \suzaku\ observation.

515: 

516: \subsection{Spectral analysis} 

517: 

518: \subsubsection{The X-ray continuum} 

519: 

520: The results of the spectral fits to the XRT and BAT spectra are

521: reported in Table~4. The best-fit parameters and their 90\%

522: uncertainties for one parameter of interest ($\Delta\chi^2$=2.7) are

523: listed, together with the observed fluxes and intrinsic

524: (absorption-corrected) luminosities. 

525: 

526: \noindent{\bf BAT:}

527: We fitted the BAT spectra with a single power law model without an

528: absorption column density, as the hard X-rays are not affected by the

529: relatively small column densities measured with the XRT (see

530: below). The BAT continuum from the first 3.5 months of the survey,

531: when the source was discovered, is fitted by a very hard power law,

532: $\Gamma \sim 0.74$. The isotropic X-ray luminosity of L$_{15-150~keV}

533: \sim 9 \times 10^{47}$ \lum\ makes J0746 one of the most luminous 

534: extragalactic sources at hard X-rays to-date.

535: 

536: The BAT continuum from the whole 9 months integration, however, is

537: described by a steeper power law, $\Gamma=1.3$. The derived flux is a

538: factor 2 lower than from the 3-months spectrum, implying that the

539: source was already fading after the first 3 months.  However, as shown

540: in the top panel of Figure~5, this is only a \gtsima 1$\sigma$

541: effect. The high-energy spectral variability properties of J0746 will

542: be better quantified in the \suzaku\ HXD observations. 

543: 

544: %To investigate this into better detail, we checked that the BAT

545: %3-months spectrum is inconsistent with the 9-months one. Superposing

546: %$\Gamma=1.3$ power law on the 3-months BAT spectrum yields a poor

547: %match, with large-amplitude residuals at the higher-energy end of the

548: %spectrum which indicate unaccounted excess flux.  Thus, we conclude

549: %that flux and spectral variability is present at energies $>$ 10 keV

550: %on timescales of months, with a trend of steeper continuum with

551: %decreasing flux. To better quantify the high-energy variability

552: %properties of J0746, future more sensitive \suzaku\ observations are

553: %urged. 

554: 

555: \noindent{\bf XRT:}

556: The XRT spectra were initially fitted with a single power law with the

557: Galactic column density. At all epochs, the residuals of this model

558: showed a deficit of flux below 3 keV, indicating the presence of

559: spectral flattening at these energies. This is illustrated in the

560: bottom panel of Figure~5, where the residuals of a power law with

561: Galactic N$_H$ from the joint fits to the 4 epochs is plotted. The

562: flattening below 3 keV can be interpreted as due either to excess

563: absorption over Galactic, or as a flatter power law component. We

564: tested both possibilities by fitting the XRT data with two models: 1)

565: a single power law plus an additional free column density, N$_H^z$, at

566: the source rest-frame; and 2) a broken power law with Galactic

567: absorption only.

568: 

569: For model 1, the choice of an intrinsic absorber is motivated by the

570: fact that no evidence for absorption due to an intervening medium is

571: present in the optical spectra of the quasar (Fig. 2). Moreover,

572: excess X-ray absorption along the line of sight was previously found

573: to be common in $z>1$ radio-loud quasars (see Worsley et al. 2006 and

574: references therein), but not in coeval radio-quiet sources, suggesting

575: the absorbing medium is located near the central engine of the AGN

576: (Elvis et al. 1992, Fiore et al. 1998, Page et al. 2005).

577: 

578: The fit with model 1 is reported in Table~4. In the three longest

579: observations (obsids 002, 003, 007), the addition of the intrinsic

580: absorber is significant at 99\% confidence ($\Delta\chi^2$=5) for 002

581: and 003, and at 92\% ($\Delta\chi^2$=4) confidence for 007, according

582: to the F-test.  The measured intrinsic column density is N$_H^z=(1-4)

583: \times 10^{22}$ \nh, and the photon index is $\Gamma=1.3$, at the flat

584: end of the distribution for $z>1$ radio-loud quasars with $\log R_L

585: \approx 3$ (Page et al. 2005). In the case of obsid 001, corresponding

586: to the lowest intensity state and shortest exposure time, the column

587: density and slope from a free-fit are poorly constrained; we thus

588: repeated the fit by fixing N$_H^z$ to the average value from the fits

589: to the remaining three observations. This yields $\Gamma \sim 1.3$,

590: similar to the other three exposures and confirming that the

591: 0.5--8~keV flux changes with no accompanying spectral variations

592: (within 20\%).

593: 

594: Extrapolating the best-fit model of the November XRT observation to

595: energies $>$ 10~keV yields a F$_{15-150~keV}=2.4 \times 10^{-11}$

596: \flux. This is a factor 4 smaller than the flux from the 9-months

597: survey, when the BAT flux was already declining. It is thus likely 

598: that during the November 2005 campaign the source was in a low state.

599: 

600: Acceptable fits were obtained from model 2, a broken power law with

601: Galactic absorption; the parameters are reported in Table~4. Indeed,

602: in the case of the longest exposure of November, the broken power law

603: model is formally preferred (at 98\% confidence, $\Delta\chi^2$=8) to

604: model 1.

605: %The broken power law fit to

606: %the November XRT spectrum yields $\Gamma_{soft}=1.04^{+0.11}_{-0.17}$

607: %and $\Gamma_{hard}=2.14^{+1.17}_{-0.66}$ below and above an observed

608: %break energy E$_b=4.0^{+0.9}_{-2.5}$~keV, respectively. 

609: This is a well-known observed phenomenon in high-z radio-loud quasars:

610: the spectral flattening at soft energies can be equally well described

611: by a power law with absorption (either cold or warm) and a convex

612: broken power law. We will revisit this issue in \S~5.3. 

613: 

614: We thus conclude that the XRT spectra are well-described by a single

615: power law with strong spectral flattening below a few keV, which can

616: be described in terms of excess absorption over the Galactic value

617: (Table~4). Acceptable fits are also obtained with a convex broken

618: power law, yielding $\Delta\Gamma \sim 1.1$ at 4~keV. 

619: 

620: \subsubsection{Limits on the \feka\ emission line}

621: 

622: Usually, in blazars the beamed jet emission dilutes or overwhelms the

623: reflection features from the disk, yielding a featureless X-ray

624: spectrum (Sambruna, Chou, \& Urry 2000). However, during a low state

625: of the non-thermal emission the \feka\ line can be detected. For

626: example, in 3C~273, Grandi \& Palumbo (2004) observed a variable EW of

627: the \feka\ line and estimated that the X-ray disk spectrum is diluted

628: by the beamed radiation of the jet by a factor 1.2 to 2.8 from 

629: 2--10~keV, while the jet always dominates above 40~keV. Similarly, a

630: \feka\ line with EW $\sim$ 100 eV was detected from the

631: cores of two powerful quasars hosting one-sided \chandra\ jets

632: (Sambruna et al. 2006, in prep.). 

633: 

634: We thus investigated the presence of an \feka\ emission line in J0746,

635: which in this source is redshifted to 1.61 keV, at the peak of the

636: detector effective area. We concentrated on the Nov 4 spectrum, which

637: has the best signal-to-noise ratio. An unresolved (width

638: $\sigma_l=0.05$ keV) Gaussian line with rest-frame energy fixed at

639: 6.4~keV was added to the power law best-fit model.  The 90\%

640: confidence upper limit to the rest-frame EW is 200 eV.

641:  

642: No evidence for such a line is present in the other datasets. Adding a

643: Gaussian line at 1.61~keV to the power law model that best-fits the

644: four datasets jointly yields $\Delta\chi^2=-1$, indicating that the

645: Gaussian component is not needed.

646: 

647: \section{Discussion} 

648: 

649: \subsection{What is J0746?} 

650: 

651: Below we discuss the evidence supporting the classification of J0746

652: as a high luminosity blazar, or more precisely a Flat Spectrum Radio

653: Quasar (FSRQ), due to its prominent broad emission lines in the

654: optical. The high energy emission however can only be attributed to a

655: relativistic jet. 

656: 

657: \subsubsection{Multiwavelength Variability} 

658: 

659: One of the defining properties of blazars is their variability, often

660: dramatic, at all observed wavelengths from radio to gamma-rays (Ulrich

661: et al. 1995). Indeed, in J0746 flux and spectral variability is

662: observed in the radio and X-rays, the two bands with the best coverage

663: so far.

664: 

665: J0746 was observed as part of the Green Bank Interferometry monitoring

666: program at 2.3 and 8~GHz during the years 1979-1996 (Lazio et

667: al. 2001). Correlated flux and spectral variability was observed at

668: 8~GHz, with a flux increase by a factor 3 in 3 years and steeper slope

669: with increasing flux. This is similar to other FSRQs in the Lazio et

670: al. sample. 

671: 

672: In the optical-UV, the \swift\ UVOT measurements are consistent with

673: the magnitudes from the SDSS, and with historical records

674: (\S~3.3). While this needs to be confirmed by future monitoring, the

675: lack of variability in this band is consistent with the idea that the

676: optical/UV radiation comes from a different component, most likely the

677: thermal emission from the disk. A straightforward test could be

678: provided by polarimetry, with a highly polarized optical flux arguing

679: against a thermal origin and in favor of a non-thermal one.  

680: 

681: J0746 was detected by the BAT at $>$ 10~keV with a bright flux. This

682: is similar to other blazars. Bright ($\sim 10^{-10}$ \flux) emission

683: was detected with the BAT (Giommi et al. 2006, submitted) and

684: \integral\ (Pian et al. 2006) from the $z$=0.859 FSRQ 3C~454.3, while

685: a rapid (lasting 2~ks) flare was detected with \integral\ in

686: 20--40~keV at the position of the $z$=0.902 blazar NRAO~530 (Foschini

687: et al. 2006). Thus, it is reasonable to expect significant

688: shorter-term flux variability at energies $>$ 10~keV from J0746 in

689: future more sensitive observations. 

690: 

691: The X-ray variability is better constrained in the 0.5--8~keV energy

692: band. The XRT light curves of J0746 show flux changes of a factor 2 on

693: timescales of hours to weeks and months. There is a hint that the hard

694: X-rays vary with larger amplitude than the softer energies, but this

695: finding needs to be confirmed by future observations. However, while

696: the 0.5--8~keV flux varies, there are no accompanying spectral

697: variations. This is similar to other FSRQs at lower $z$ observed with

698: \asca\ and \rosat\ (Donato et al. 2001; Sambruna 1997), and to other

699: radio-loud quasars at $z>1$ (Page et al. 2005). 

700: 

701: The large X-ray luminosity and rapid variability of J0746 provide

702: further support for the idea that its emission is beamed. In fact,

703: assuming the X-ray emission from J0746 is isotropic, we can derive a

704: limit to the radiative efficiency $\eta$. Following Fabian (1979),

705: $\eta > 4.8 \times 10^{-43} \Delta L/\Delta t$ erg s$^{-1}$ s$^{-1}$,

706: where $\Delta L$ is the luminosity change in a time interval $\Delta

707: t$ (Brandt et al. 1999). From Table~4 and Figure 4d, $\Delta L \sim 5

708: \times 10^{46}$ \lum\ and $\Delta t \sim 1800$ sec in the quasar's

709: rest-frame. Thus, $\eta > 13$, which is unphysically large unless we

710: require relativistic beaming of the radiation.

711: 

712: \subsubsection{Broad-band spectral indices} 

713: 

714: The broad-band energy distributions of blazars can be described to 

715: first approximation with the spectral indices $\alpha_{ro}$,

716: $\alpha_{ox}$, and $\alpha_{rx}$. These are defined as the

717: two-point indices between radio (5~GHz) and optical ($V$ band),

718: optical and X-rays (1~keV), and radio to X-rays. Based on the

719: radio-to-X-ray flux ratio, blazars can be further classified as

720: High-Energy Peaked BL Lacs (HBLs), with $\alpha_{rx} < 0.8$; and

721: Low-energy Peaked BL Lacs (LBLs) and FSRQs, with $\alpha_{rx} > 0.8$

722: (Padovani \& Giommi 1995). 

723: 

724: When plotted in the $\alpha_{ro}- \alpha_{ox}$ plane, blazar classes

725: occupy distinct, although partly overlapping, regions. For HBLs,

726: $\alpha_{ro} < 0.6$ and $\alpha_{ox} < 1.3$, while LBLs and FSRQs have

727: $\alpha_{ro} > 0.5$ and $\alpha_{ox} > 1.0$ (e.g., Donato et

728: al. 2001). Using rest-frame flux densities at 5~GHz, in $V$ band, and

729: at 1~keV, the broad-band spectral indices of J0746 are

730: $\alpha_{rx}=0.78, \alpha_{ro}=0.65$, and $\alpha_{ox}=1.03$. Thus,

731: J0746 appears to be a borderline source between FSRQs and HBLs.

732: 

733: It is useful to compare the broad-band indices of J0746 to radio-quiet

734: quasars, because this could indicate the level of boosting of the

735: continuum due to beaming. A recent extensive compilation of broad-band

736: indices for radio-quiet quasars is provided by Strateva et

737: al. (2005). Here, they define an optical-to-X-ray index as the slope

738: between the flux at 2~keV and at 2500~\AA\ in the quasar's

739: rest-frame. For J0746, this index is 1.16, much flatter than the

740: radio-quiet quasars in Strateva et al. (2005). Thus, for a given

741: optical flux, the X-ray emission from J0746 is more than a factor 10

742: larger than expected for radio-quiet quasars. This also suggests that,

743: at least to a first approximation, beaming affects the X-ray emission

744: of J0746 but not (or much less) its optical-UV flux.

745: 

746: \subsubsection{The Spectral Energy Distribution (SED)} 

747: 

748: As discussed in \S~1, a strong spectral signature of blazars is the

749: presence of a double-peaked structure in their SEDs. Similarly, as

750: shown in Figure~6, J0746 exhibits a double-humped SED. 

751: 

752: To assemble the SED of J0746, we used data from this paper as well as

753: from the literature. In Figure~6 the radio (filled circles), optical,

754: UV, and X-ray (XRT, 0.5--8~keV) fluxes are contemporaneous

755: measurements from our November 4 campaign, while the filled triangles

756: are archival observations from \verb+NED+. Also plotted is the

757: 9-months BAT spectrum (see below). As such, the SED in Figure~6

758: contains contemporaneous (radio, optical, UV, and medium-hard X-rays)

759: data as well as archival fluxes.

760: 

761: At GeV energies, we plot an upper limit to the EGRET flux from a

762: reanalysis of four archival images. Unfortunately, the coverage of the

763: J0746 field by EGRET was very poor: EGRET observed the source in 1992,

764: 1993, 1994, and 1995. However, J0746 was never closer than 16.7\deg\

765: to the EGRET axis; this implies that no high-quality exposure of the

766: source was ever acquired. The upper limit to the $>$ 100 MeV flux from

767: the combined four observations is $5\times 10^{-8}$ ph cm$^{-2}$

768: s$^{-1}$ and is plotted in Figure~6 with an arrow.

769: 

770: The first striking feature in Figure~6 is that, as implied by the hard

771: BAT spectrum and the EGRET limit, the peak of the second spectral

772: component is located around a few MeV. This would make J0746 a new

773: member of the poorly known MeV blazars subclass, which so far contains

774: only two sources detected by COMPTEL (Blom et al. 1996; for PKS

775: 0208--512 see, however, Stacy et al. 2003). The obvious caveat is, of

776: course, that the BAT and EGRET fluxes are not simultaneous, while

777: FSRQs notoriously vary at high energies on timescales of days.

778: 

779: The second important feature in Figure~6 is the steep optical-to-UV

780: continuum, which implies a peak of the synchrotron component at IR

781: wavelengths. Observations of J0746 in the near and far-IR are urged to

782: better sample the SED in this critical portion of the spectrum. It is

783: tempting to interpret the steep optical-UV emission as the high-energy

784: tail of thermal emission from the accretion disk, i.e., the so-called

785: blue bump, usually observed in quasars (Elvis et al. 1994) but also in

786: a few FSRQs (3C~273; von Montigny et al. 1997).

787: 

788: In conclusion, based on multiwavelength evidence, we classify J0746

789: with a high-$z$ blazar. Further observations, especially polarimetry

790: at radio through optical wavelengths, are strongly encouraged to

791: confirm the blazar nature of this source. 

792: 

793: Below we model the SED of J0746 in the assumption that J0746 is a

794: blazar, and that the optical-to-UV continuum at the time of the UVOT

795: observations is due to thermal emission from the disk.

796: 

797: \subsection{Modeling the Spectral Energy Distribution} 

798: 

799: The continuum from blazars is generally attributed to synchrotron and

800: inverse-Compton (IC) emission (e.g., Ghisellini et al. 1998; for a

801: different view see Atoyan \& Dermer 2003 and references therein)

802: produced within a jet pointing toward the observer. In the case of

803: blazars showing strong emission lines, as in J0746, the target photons

804: for the IC process are likely dominated by those belonging to the

805: external radiation field (disk, broad line region), amplified by

806: relativistic effects in the plasma rest-frame (Dermer \& Schlickeiser

807: 1993, Sikora et al. 1994). Synchrotron self-Compton emission (Maraschi

808: et al. 1992), on the other hand, can substantally contribute in the

809: soft X-ray band (e.g., Ballo et al. 2002).

810: 

811: Again, we caution that the SED in Figure~6 contains non-simultaneous

812: data. In particular, the BAT and XRT spectra were taken several months

813: apart, while the BAT spectrum was integrated over a long-time period

814: (9 months). The choice to use the 9-months BAT spectrum in the SED

815: modeling is due to the fact that it is ``closer'' in time to the Nov 4

816: campaign data; indeed, the extremely flat continuum during the initial

817: higher state, $\alpha \sim -0.26$, posits severe problems for current

818: blazar models.

819: 

820: To reproduce the observed SED we have applied the synchrotron-IC

821: emission model fully described in Maraschi \& Tavecchio

822: (2003). Briefly, the emission region is modeled as a sphere with

823: radius $R$, in motion with bulk Lorentz factor $\Gamma$ at an angle

824: $\theta $ with respect to the line of sight, and filled with tangled

825: magnetic field (with intensity $B$) and relativistic

826: electrons. $\Gamma$ and $\theta$ are combined in the Doppler factor

827: $\delta=[\Gamma (1-\beta \cos \theta)]^{-1}$, where $\beta=v/c$.  The

828: (purely phenomenological) electron distribution is modeled as a

829: smoothed broken power law, with indices $n_1$ and $n_2$ below and

830: above the break Lorentz factor $\gamma_b$. The electron distribution

831: extends within the limits $\gamma_{\rm min} <\gamma <\gamma_{\rm

832: max}$.

833: 

834: The exceptionally steep optical-UV continuum strongly suggests that

835: the emission in this band is dominated by a thermal-like component,

836: possibly associated with the blue-bump component which characterizes

837: the quasar optical region (e.g., Sun \& Malkan 1989), usually thought

838: to be associated with the putative accretion disk. This interpretation

839: is justified by the presence of prominent emission lines in the HET

840: spectrum of J0746 (\S~3.2), which was taken a few weeks after the end

841: of the BAT observations discussed here. 

842: 

843: The identification of the optical-UV component with the disk emission

844: constrains the disk luminosity L$_D \sim$ a few $\times 10^{47}$

845: \lum. We choose to model this component with a black body peaking at

846: the typical frequency of $10^{15}$ Hz (in the quasar rest-frame). The

847: external radiation field is modelled as the same black-body, diluted

848: within the broad line region, supposed to be spherical with radius

849: $R_{\rm BLR}$. This radiation approximates the radiation reprocessed

850: and re-isotropized by the BLR clouds. Its luminosity is fixed to a

851: fraction $\tau $ of the luminosity $L_D$ of the putative central

852: illuminating disk.

853: 

854: The resulting model is reported in Figure~6 (solid line), with the

855: parameters listed in the caption. From the inferred parameters and

856: assuming the composition of one proton per emitting electron (for the

857: reason of this choice see Maraschi \& Tavecchio 2003), we can infer a

858: kinetic power of the jet of $P_j=4\times 10^{48}$ \lum. This can be

859: compared with the radiative output of the disk $L_{\rm D} = 4\times

860: 10^{47}$ \lum. The ratio $P_j/L_D\sim 10$ is of the same order of what

861: is found in other cases for which both powers can be estimated with

862: reasonable accuracy (Tavecchio et al. 2000; Maraschi \& Tavecchio

863: 2003). The parameters used to reproduce the SED are close to those

864: usually found for other powerful blazars (e.g., Maraschi \& Tavecchio

865: 2003, Ghisellini et al. 1998).

866: 

867: An interesting aspect of the \swift\ results is the exceptionally flat

868: spectrum measured in the hard X-rays. Flat spectra ($\alpha_X = 0.3$)

869: characterize several high-luminosity blazars (e.g., Tavecchio et

870: al. 2000). To model such hard spectra we use an electron index of

871: $n_1=1.5$. Such a flat electron distribution poses serious problems

872: for the standard scenarios usually discussed, involving shock

873: acceleration or cooling processes, which lead to a typical

874: distribution with $n=2$. An interesting possibility has been

875: considered by Sikora et al. (2002). They assume a two-step

876: acceleration process: the flat portion is produced by a

877: pre-acceleration mechanism (e.g., stochastic acceleration driven by

878: turbulence or reconnection), while the steep portion after the peak is

879: produced through the standard shock acceleration.

880: 

881: In the radio band, our data fits well to a power law spectrum with

882: $\alpha=0.48$, quite typical of a transparent synchrotron source.

883: Without a high resolution VLBI image, the most straight forward

884: interpretation is that there is an extended region around the central

885: core that is responsible for the radio-mm flux.  The IR-optical

886: emission is presumably produced in a core region that starts appearing

887: at shorter wavelengths, but without data in the sub-mm it is not

888: possible to specify exactly where the BLR region starts to dominate.

889: The discontinuity between the radio-mm spectrum and the shorter

890: wavelength emission is consistent with a high energy cutoff in the

891: radiating particles in this extended emitting region; such a cutoff is

892: already part of the model parameters used to fit the core region at

893: shorter wavelengths.

894: 

895: It is worth noting that we fix $\gamma_{\rm min}$ to 1. Larger

896: values would in fact produce an unobserved deficit of soft X-ray

897: photons, further flattening the X-ray spectrum. This is also typical

898: of this kind of sources, and seems to support the two-step view, in

899: which the bulk of the electrons remains at low energies.

900: 

901: Finally, while we have assumed that the steep optical-UV continuum in

902: Figure~6 is thermal in origin and related to the accretion disk, we

903: can not exclude {\it a priori} that the jet does not contribute to a

904: fraction of the optical flux, or even dominates it during

905: outbursts. In this sense, variability of the optical-UV continuum in

906: J0746 would have to be modeled as a mix of a non-thermal, variable jet

907: component and a thermal, less variable one related to the accretion.

908: 

909: \subsection{J0746 in context} 

910: 

911: There is evidence that the double-humped SEDs of blazars form a

912: continuous sequence in luminosity (Sambruna, Maraschi, \& Urry

913: 1996). In this sequence, going from the more luminous FSRQs to the

914: fainter and closer TeV BL Lacs, the synchrotron component peak shifts

915: from IR to X-ray wavelengths, while the Compton peak shifts, by the

916: same amount, from GeV to TeV gamma-rays (Fossati et al. 1998). Models

917: of the SED sequence suggest that particle energies decrease and

918: magnetic fields increase with increasing luminosities while the

919: Doppler factor is approximately constant (Ghisellini et al. 1998). The

920: sequence finds a physical basis in the run of the jet and disk powers,

921: P$_{jet}$ and P$_{disk}$, with P$_{jet} \sim$ P$_{disk} \sim

922: 10^{46}-10^{47}$ \lum\ for FSRQs and P$_{jet} >$ P$_{disk}$ for BL

923: Lacs (Tavecchio et al. 2000; Maraschi \& Tavecchio 2003). 

924: 

925: %In this scenario, blazar jet SEDs

926: %represent an evolutionary scenario governed by the accretion rate

927: %(B\"ottcher \& Dermer 2002), with FSRQs accreting at the Eddington

928: %limit and BL Lacs at sub-Eddington rates (Maraschi \& Tavecchio 2003).

929: 

930: To compare the SED of J0746 to other blazars, we have plotted in

931: Figure~7 the blazar luminosity sequence from the reanalysis of Donato

932: et al. (2001). The solid line represents the best-fit model to the SED

933: of J0746 from Figure~6. It is apparent that the spectral properties of

934: this source are extreme in the blazar class: as discussed above, the

935: synchrotron peak likely lies below the IR frequency range while the IC

936: peak is at MeV energies. This qualifies J0746 as a member of the

937: poorly known ``MeV blazars'' subclass (Blom et al. 1996).

938: 

939: There is much debate whether the blazar sequence is intrinsically true

940: or a product of several observational biases, especially at gamma-rays

941: (Urry 1999), or limited sensitivity of current radio and optical

942: surveys (Nieppola, Tornikoski, \& Valtaoja 2006; Anton et

943: al. 2004). Nevertheless, here we used the sequence as a comparison

944: template to showcase the unusual properties of J0746. Note that, if

945: the sequence is true, the SED of J0746 would be extreme, but not

946: unexpected. 

947: 

948: A property J0746 shares with other high-$z$ radio-loud quasars is the

949: presence of spectral hardening in the X-ray spectrum below a few

950: keV. As discussed in \S~4.2.1, this can be modeled in terms of either

951: excess absorption along the line of sight, or a flatter power law. We

952: now discuss the two scenarios in turn. 

953: 

954: In the context of blazar emission models, a continuum break at soft

955: X-rays can be expected if the particle distributions responsible for

956: the high-energy component have a low-energy cutoff, parameterized by

957: the value of the minimum Lorentz factor of the relativistic electrons

958: (Tavecchio et al. 2000). In this scenario, the hardening occurs below

959: $\sim \gamma_{min}^2 \Gamma^2 \nu_{ext}$, with $\nu_{ext} \sim

960: 10^{15}$ Hz (Fig. 6). The fitted value of the break energy in J0746

961: would thus imply $\gamma_{min} \sim 3$, consistent with the SED

962: modeling (Fig. 6). 

963: %A possible difficulty for this model is that such

964: %extreme breaks as observed in J0746 are rarely seen in other quasars,

965: %especially at lower-$z$.  

966: 

967: %Alternatively, the flattening can be produced by a

968: %sharply peaked external distribution of seed photons (Fabian et

969: %al. 2001). In this scenario, the external radiation density is

970: %strongly boosted by a factor $\Gamma^2$ in the observer's frame,

971: %implying a decay of the flux below $\Gamma^2 \nu_{ext}$ (Fabian et

972: %al. 2001). Plugging in the value of the break energy for J0746, this

973: %would yield $\Gamma \sim 40$, a factor 2 within the value derived from

974: %the SED (Fig. 6). 

975: 

976: Alternatively, the spectral flattening below 3 keV can be described in

977: terms of excess N$_H$ along the line of sight, as originally advocated

978: for high-z radio-loud quasars by Elvis et al. (1992). Recent \xmm\

979: observations have confirmed that at least a fraction of these sources

980: could have intrinsic absorption, with N$_H^z \sim 10^{22}-10^{23}$

981: \nh\ (see Worsley et al. 2006 and references therein). The nature and

982: location of the absorber, however, is still debated today. The large

983: X-ray column densities would imply large optical reddening for the

984: quasars, which is not seen.

985: 

986: Such difficulty exists for J0746 as well. The SDSS and HET spectra

987: indeed show little or no intrinsic reddening, at odds with the X-ray

988: column density N$_H \sim 10^{22}$ \nh\ which implies $A_V \sim 10$ 

989: (assuming Galactic dust-to-gas ratios). Moreover, while several X-ray

990: absorbed quasars, both radio-loud and radio-quiet, show accompanying

991: UV absorption lines (e.g., Brandt, Laor, \& Wills 2000), this is not

992: the case for J0746 (Fig. 2). This would seem to indicate a dust-free,

993: neutral absorber. However, perhaps the major difficulty is to devise a

994: physically plausible location for the absorber. As shown in Figure~6

995: the X-ray emission from J0746 comes from a highly beamed ($\delta \sim

996: 20$) jet, at least in its inner regions. One can thus speculate that

997: the absorbing medium is located inside the jet, perhaps ambient gas

998: entangled in the moving blobs (e.g., Celotti et al. 1998). 

999: %There is indeed evidence that

1000: %entrainment of ambient gas can yield enough mass-loading to decelerate

1001: %even powerful jets on kpc-scales (Tavecchio et al. 2006); 

1002: However, whether this gas has a covering factor large enough to

1003: produce significant spectral depression in the X-ray continuum

1004: (Fig.~5) remains to be proven.

1005: 

1006: In conclusion, the nature of the low-energy spectral flattening in

1007: J0746 is still an open question. We look forward to the deeper 

1008: (100~ks), broad-band \suzaku\ spectrum which will hopefully settle the

1009: issue.

1010: 

1011: \section{Summary and Conclusions} 

1012: 

1013: The BAT experiment onboard \swift\ detected bright emission in the

1014: energy range 15--150~keV from the radio-loud quasar J0746 at

1015: $z$=2.979, implying an isotropic luminosity around $\sim 10^{48}$

1016: \lum. Rapid variability in 0.5--10~keV strongly suggests beaming. 

1017: In this paper, we reported the \swift\ discovery observations and

1018: follow-up multiwavelength monitoring of this source. The main results

1019: are:

1020: 

1021: \begin{itemize} 

1022: 

1023: \item The X-ray emission of J0746 is variable on timescales 

1024: of days to months (Fig. 3, 4). Fast variability on timescales of

1025: 1--2~ks is observed for the 0.5--8~keV flux (Fig.~3b). 

1026: 

1027: \noindent At all epochs, the XRT hardness ratio light curve is

1028: consistent with a constant value, implying there is no spectral

1029: variability on short (hours) timescales in 0.5--8~keV.

1030: 

1031: \item The 15--150~keV emission is consistent with a very hard 

1032: continuum, $\Gamma \sim 0.74$, during the first 3 months of the BAT

1033: survey, and a steeper continuum, $\Gamma \sim 1.3$, with halved flux, 

1034: in the BAT spectrum integrated over 9-months. 

1035: 

1036: \item The 0.5--8~keV spectrum is well described by a $\Gamma \sim 1.3$

1037: power law. Spectral flattening is present at $<$ 1 keV, which can be

1038: interpreted as excess absorption over the Galactic value, implying an

1039: intrinsic column density N$_H^z \sim 10^{22}$ \nh, or a flatter power

1040: law component, implying a sharp ($\Delta \Gamma \sim 1.1$) spectral

1041: break around 16~keV in the quasar's rest-frame.

1042: 

1043: \item The Spectral Energy Distribution of J0746 has a 

1044: double-humped structure, with the first component peaking around IR

1045: wavelengths and the second component likely at a few MeV. The

1046: luminosity of the source is dominated by the high-energy emission. 

1047: 

1048: \end{itemize} 

1049: 

1050: Based on the results presented in this paper, we conclude that J0746

1051: is a blazar, and more precisely, a Flat Spectrum Radio Quasar. Its SED

1052: has extreme, but not unexpected, properties compared to other

1053: blazars. Additional and possibly simultaneous multiwavelength

1054: observations of J0746 are encouraged to confirm the jet-origin of its

1055: emission; to this end, polarization measurements at radio and

1056: optical wavelengths will be especially revealing.

1057: 

1058: The detection of a flare at high-energies from J0746 raises several

1059: outstanding questions. How common are sources like J0746? What are the

1060: physical parameters of MeV blazar jets? What causes the violent

1061: outbursts at hard X-rays, and what is their duty cycle? The latter

1062: issue is particularly important in view of the possibility that

1063: blazars are significant, and even dominant, contributors to the

1064: gamma-ray sky (Salamon \& Stecker 1994). The BAT flux variability on

1065: timescales of months, as well as the BAT detection of several other

1066: sources with properties similar to J0746 (Sambruna et al., in prep.),

1067: suggests that flares at energies $>$ 10 keV may be common from these

1068: blazars, stressing the key importance of wide-angle, sensitive and

1069: continuous sky monitoring at these energies. 

1070: 

1071: \acknowledgements

1072: 

1073: This research has made use of data obtained from the High Energy

1074: Astrophysics Science Archive Research Center (HEASARC), provided by

1075: NASA's Goddard Space Flight Center, and of the NASA/IPAC Extragalactic

1076: Database (NED) which is operated by the Jet Propulsion Laboratory,

1077: California Institute of Technology, under contract with the National

1078: Aeronautics and Space Administration. UMRAO is partially funded by a

1079: series of grants from NSF and NASA and by the University of

1080: Michigan. This work was supported in part by National Science

1081: Foundation grant AST-0307582 (DPS) and NASA LTSA grant NAG5-13035

1082: (WNB, DPS). 

1083: 

1084: Funding for the creation and distribution of the SDSS Archive has been

1085: provided by the Alfred P. Sloan Foundation, the Participating

1086: Institutions, the National Aeronautics and Space Administration, the

1087: National Science Foundation, the U.S.  Department of Energy, the

1088: Japanese Monbukagakusho, and the Max Planck Society.  The SDSS Web

1089: site \hbox{is {\tt http://www.sdss.org/}.}  The SDSS is managed by the

1090: Astrophysical Research Consortium (ARC) for the Participating

1091: Institutions.

1092: 

1093: The Hobby-Eberly Telescope (HET) is a joint project of the University

1094: of Texas at Austin, the Pennsylvania State University, Stanford

1095: University, Ludwig-Maximillians-Universit\"at M\"unchen, and

1096: Georg-August-Universit\"at G\"ottingen.  The HET is named in honor of

1097: its principal benefactors, William P. Hobby and Robert E. Eberly.  The

1098: Marcario Low-Resolution Spectrograph is named for Mike Marcario of

1099: High Lonesome Optics, who fabricated several optics for the instrument

1100: but died before its completion; it is a joint project of the

1101: Hobby-Eberly Telescope partnership and the Instituto de

1102: Astronom\'{\i}a de la Universidad Nacional Aut\'onoma de M\'exico. 

1103: 

1104: \begin{references}

1105: 

1106: \reference{} Aller, H. D., Aller, M. F., Latimer, G. E., \& Hodge,

1107:              P. 1985, ApJS, 59, 513

1108: \reference{} Anton, S., Browne, I. W. A., Marchã, M. J. M., Bondi, M.,

1109:              \& Polatidis, A. 2004, MNRAS, 352, 673 

1110: \reference{} Atoyan, A.~M., \& Dermer, C.~D. 2003, ApJ, 586, 79

1111: \reference{} Baars, J. W. M., Genzel, R., Pauliny-Toth, I. I. K., 

1112:           \& Witzel, A. 1977, A\&A, 61, 99 

1113: \reference{} Ballo, L., et al.\ 2002, ApJ, 567, 50

1114: \reference{} Barthelmy, S.D. et al. 2005, Space Science Rev., in

1115:              press (astroph/0507410)

1116: \reference{} Bauer, F. et al. 2004, AJ, 128, 2048

1117: %\reference{} Bennett, C.L. et al. 2003, ApJS, 148, 1

1118: \reference{} Blom, J.J. et al. 1996, A\&AS, 120, 107 

1119: %\reference{} B\"ottcher \& Dermer 2002, ApJ, 564, 86

1120: \reference{} Brandt, W.N., Laor, A., Wills, B.J. 2000, ApJ, 528, 637 

1121: \reference{} Brandt, W.N., Boller, Th, Fabian, A.C. \& Ruszkowski,

1122:       M. 1999, MNRAS, 303, L53 

1123: \reference{} Capalbi, M., Perri, M., Saijia, B., Tamburelli, F., \& Angelini, L. 2005, 

1124:       ``The SWIFT XRT Data Reduction Guide'', version 1.2 

1125: \reference{} Cappi, M. et al. 1997, ApJ, 478, 492 

1126: \reference{} Celotti, A., Kuncic, Z., Rees, M.~J., \& Wardle, J.~F.~C. 

1127:              1998, MNRAS, 293, 288 

1128: \reference{} Dermer, C.D. \& Schlickeiser, R. 1993, ApJ, 416, 458

1129: \reference{} Donato, D., Ghisellini, G., Tagliaferri, G., \& Fossati,

1130:              G. 2001, A\&A, 735, 739

1131: \reference{} Edelson, R., Turner, T. J., Pounds, K., Vaughan, S., 

1132:        Markowitz, A., Marshall, H., Dobbie, P., \& Warwick, R. 2002, ApJ, 568, 61

1133: \reference{} Elvis, M. et al. 1994, ApJS, 95, 1 

1134: \reference{} Elvis, M., Fiore, F., Wilkes, B., McDowell, J., \&

1135:          Bechtold, J. 1992, ApJ, 422, 60 

1136: \reference{} Fabian, A.C. 1979, Proc. R. Soc. London, Ser. A., 366, 449

1137: \reference {} Fabian, A.C. et al. 2001, MNRAS, 323, 373 

1138: %\reference{} Fiore, F., Elvis, M., Giommi, P. \& Padovani, P. 1998,

1139: %             ApJ, 492, 79 

1140: \reference{} Fiore, F., Elvis, M., Giommi, P., \& Padovani, P. 1998,

1141:              ApJ, 492, 79 

1142: \reference{} Foschini, L. et al. 2006, A\&A, in press (astroph/0601101)

1143: \reference{} Fossati, G., Maraschi, L., Celotti, A., Comastri, A., \&

1144:          Ghisellini, G. 1998, MNRAS, 299, 433 

1145: \reference{} Fukugita, M., Ichikawa, T., Gunn, J.E., Doi, M., Shimasaku, K.,

1146:      \& Schneider, D.P. 1996, AJ, 111, 1748

1147: \reference{} Ghisellini, G. et al. 1998, MNRAS, 301, 451

1148: \reference{} Giommi, P. et al. 2006, A\&A, submitted

1149: \reference{} Grandi, P. \& Palumbo, G.C. 2004, Science, 306, 998 

1150: \reference{} Hill, G.J., Nicklas, H.E., MacQueen, P.J., Mitsch, W., Wellem, W.,

1151:      Altmann, W., Wesley, G.L., \& Ray, F.B. 1998, Proc. SPIE, 3355, 433

1152: \reference{} Laurent-Muehleisen, S. A., Kollgaard, R. I., Ryan, P. J.,

1153: Feigelson, E. D., Brinkmann, W., \& Siebert, J. 1997, A\&AS, 122, 235

1154: \reference{} Lazio, T.J.W., Waltman, E. B., Ghigo, F. D., Fiedler,

1155:     R. L., Foster, R. S., \& Johnston, K. J. 2001, ApJS, 136, 265 

1156: \reference{} Maraschi, L., \& Tavecchio, F. 2003, ApJ, 593, 667

1157: \reference{} Maraschi, L., Ghisellini, G., \& Celotti, A. 1992, ApJ, 397, L5

1158: \reference{} Markwardt, C. et al. 2006, ApJ, in press 

1159: \reference{} Nieppola, E., Tornikoski, M., \& Valtaoja, E. 2006, A\&A, 445, 441

1160: \reference{} Page, K. L., Reeves, J. N., O'Brien, P. T., \& 

1161:              Turner, M. J. L. 2005, MNRAS, 364, 195

1162: \reference{} Perlman, E.S., Padovani, P., Giommi, P., Sambruna, R.M., 

1163:           Jones, L. R., Tzioumis, A., \& Reynolds, J. 1998, AJ, 115, 1253

1164: \reference{} Pian, E. et al. 2006, A\&A, in press (astroph/0602268) 

1165: \reference{} Ramsey, L.W., Sebring, T.A., \& Sneden, C. 1994, Proc. SPIE, 2199, 31

1166: \reference{} Richards, G.T. et al. 2006, AJ, subm. (astro-ph/0601434)

1167: \reference{} Romani, R. W., Sowards-Emmerd, D., Greenhill, L., \&

1168:              Michelson, P. 2004, ApJ, 610, L9

1169: \reference{} Salamon, M.H. \& Stecker, F.W. 1994, ApJ, 430, L21 

1170: \reference{} Sambruna, R.M. 1997, ApJ, 487, 536

1171: \reference{} Sambruna, R.M., Maraschi, L., \& Urry, C.M. 1996, ApJ, 463, 444 

1172: \reference{} Schneider, D.P., et al. 2005, AJ, 130, 367 

1173: \reference{} Sikora, M., B{\l}a{\.z}ejowski, M., Moderski, R., \&

1174:     Madejski, G.~M. 2002, ApJ, 577, 78 

1175: \reference{} Spergel, D.N. et al. 2003, ApJS, 148, 175 

1176: \reference{} Stacy, J.G., Vestrand, W.T., \& Sreekumar, P.G. 2003, ApJ, 598, 216 

1177: \reference{} Strateva, I.V., Brandt, W. N., Schneider, D. P., Vanden

1178:      Berk, D. G., \& Vignali, C. 2005, AJ, 130, 387 

1179: \reference{} Sun, W.-H., \& Malkan, M.~A. 1989, ApJ, 346, 68

1180: %\reference{} Tavecchio, F., Maraschi, L., Sambruna, R.M., Gliozzi, M.,

1181: %   Cheung, C.C., Wardle, J.F.C., \& Urry, C.M. 2006, ApJ, in press

1182: %   (astro-ph/0512389)

1183: \reference{} Tavecchio, F., et al. 2000, ApJ, 543, 535

1184: \reference{} Teräsranta, H., Tornikoski, M., Mujunen, A. et al. 1998,

1185:               A\&AS, 132, 305

1186: \reference{} Ter\"asranta et al. (1998), A\&A, 132, 305

1187: \reference{} Ulrich, M.-H., Maraschi, L., \& Urry, C.M. 1997, ARA\&A, 35, 445

1188: \reference{} Urry, C.M. 1999, Astroparticle Physics, Vol. 11, Nos. 1--2,

1189:              p. 159 

1190: \reference{} Vanden Berk, D.E., et al. 2001, AJ, 122, 549

1191: \reference{} von Montigny, C. et al. 1997, ApJ, 483, 161

1192: \reference{} Wills, B.J., Brotherton, M. S., Fang, D., Steidel, C.C., 

1193:               \& Sargent, W.L.W. 1993, ApJ, 415, 563 

1194: \reference {} Worsley, M.A., Fabian, A.C., Pooley, G.G., \& Chandler,

1195:      C.J. 2006, MNRAS, in press (astroph/0602355) 

1196: \reference{} York, D.G., Adelman, J., Anderson, J.E., Anderson, S.F., 

1197:    et al. 2000, AJ, 120, 1579

1198: 

1199: \end{references} 

1200: 

1201: 

1202: 

1203: %\vspace{2cm} 

1204: 

1205: \input{tab1}

1206: \clearpage

1207: \input{tab2}

1208: \clearpage

1209: \input{tab3}

1210: \clearpage

1211: \input{tab4}

1212: %\end{document}

1213: \newpage

1214: 

1215: 

1216: 

1217: %%%%%%%%%%%%%%%%%%% figure 1: BAT image %%%%%%%%%%%%%

1218: 

1219: \vspace{-1.0cm}

1220: \begin{figure}[h]

1221: \begin{center}

1222: \hbox{

1223: % un-comment the following line to include your fig1a.ps postscript file

1224: \psfig{figure=f1.ps,height=11cm,width=10cm}

1225: % un-comment the following line to include your fig1b.ps postscript file

1226: }

1227: \end{center}

1228: \vspace{-1.0cm}

1229: \caption{

1230: {The BAT sensitivity image of J0746, ranging from $2\sigma$ (white)

1231: to $7.5\sigma$ (black) for a total exposure time of 755~ks. 

1232: The image is 10\deg x 10\deg, with 5\arcmin\

1233: pixels.  The circle marks the position of J0746.  The orientation 

1234: is arbitrary. J0746 is detected at $7.5\sigma$. 

1235: }}

1236: 

1237: \end{figure}

1238: 

1239: \newpage 

1240: 

1241: 

1242: 

1243: %%%%%%%%%%%%%%%%%%% figure 2: HET+ SDSS SPECTRUM %%%%%%%%%%%%%

1244: 

1245: \vspace{-1.0cm}

1246: \begin{figure}[h]

1247: \begin{center}

1248: \hbox{

1249: % un-comment the following line to include your fig1a.ps postscript file

1250: \psfig{figure=f2a.ps,height=10cm,width=9cm}

1251: % un-comment the following line to include your fig1b.ps postscript file

1252: }

1253: \vspace{.2cm}

1254: \hbox{

1255: \psfig{figure=f2b.ps,height=10cm,width=9cm}

1256: }

1257: \end{center}

1258: \vspace{-1.7cm}

1259: \caption{\scriptsize

1260: {The SDSS {\bf (Top)} and HET {\bf (Bottom)} spectra of J0746, showing

1261: typical broad UV lines of a high-redshift quasar. The SDSS spectrum

1262: was taken December 19, 2001 and has a resolution R=1900; the HET

1263: spectrum, taken on October 10, 2005, has a resolution R=460.}}

1264: 

1265: \end{figure}

1266: \normalsize

1267: 

1268: \newpage 

1269: 

1270: 

1271: %%%%%%%%%%%%%%%% figure 3: BAT and average XRT light curve *******************

1272: 

1273: 

1274: \begin{figure}[h]

1275: \begin{center}

1276: \hbox{

1277: % un-comment the following line to include your fig1a.ps postscript file

1278: \psfig{figure=f3a.ps,height=7.0cm,width=7.0cm}

1279: % un-comment the following line to include your fig1b.ps postscript file

1280: 

1281: %\psfig{figure=excellent_plot.ps,height=5.5cm,width=6.0cm,angle=-90}

1282: 

1283: \vspace{1.0cm}

1284: \psfig{figure=f3b.ps,height=7.0cm,width=7.0cm}

1285: }

1286: \end{center}

1287: \vspace{-1.0cm}

1288: \caption{

1289: {{\bf (a, Left):} The BAT light curve of J0746 from 15--195~keV. Bins

1290: are 15-day averages. The start date is December 15, 2004. The

1291: vertical bars in September mark the times of the follow-up XRT

1292: observations (Table~1). {\bf

1293: (b, Right):} The \swift\ XRT light curve, obtained by averaging the

1294: count rate for each observing epoch (Table~1). Variation of the 0.5--8

1295: keV flux of a factor \gtsima 2 in 5 days are apparent. }}

1296: 

1297: \end{figure}

1298: 

1299: 

1300: \newpage 

1301: 

1302: 

1303: %%%%%%%%%%%%%%%%%%%%%% figure 4:  XRT light curve *******************

1304: 

1305: 

1306: \begin{figure}[h]

1307: \begin{center}

1308: 

1309: \hbox{

1310: % un-comment the following line to include your fig1a.ps postscript file

1311: %\psfig{figure=lc-mergeB.ps,height=14.0cm,width=16.0cm}

1312: % un-comment the following line to include your fig1b.ps postscript file

1313: 

1314: %\psfig{figure=excellent_plot.ps,height=5.5cm,width=6.0cm,angle=-90}

1315: 

1316: \vspace{1.0cm}

1317: \psfig{figure=f4.ps,height=14.0cm,width=17.0cm}

1318: }

1319: \end{center}

1320: \vspace{-1.0cm}

1321: \caption{

1322: {The \swift\ XRT light curves in total band (0.5--8~keV) at the four

1323: epochs. Significant flux variability is present in the Nov 4 and Sept

1324: 5 observations. The flux changed by a factor 2.5 in one hour

1325: in September and by  a factor

1326: \gtsima 2 in 10 days in November. There is no accompanying spectral 

1327: variability as indicated by the hardness ratio light curve which is

1328: consistent with a constant (see text). 

1329: }}

1330: 

1331: \end{figure}

1332: 

1333: 

1334: \newpage 

1335: 

1336: %%%%%%%%%%%%%%%%%%%%%%%%%% figure 5: BAT spectra + XRT residuals %%%%%%%%%%

1337: 

1338: \begin{figure}[h]

1339: \begin{center}

1340: \hbox{

1341: % un-comment the following line to include your fig1a.ps postscript file

1342: \psfig{figure=f5a.ps,height=7.0cm,width=9.0cm,angle=-90}

1343: % un-comment the following line to include your fig1b.ps postscript file

1344: }

1345: \vspace{.3cm}

1346: \hbox{

1347: \psfig{figure=f5b.ps,height=7.0cm,width=9.0cm,angle=-90}

1348: }

1349: \end{center}

1350: \vspace{-1.0cm}

1351: \caption{

1352: {\bf Top:} The BAT spectra from the 3 months {\it (open triangles)} and 9

1353: months {\it (filled triangles)} integrations. There is a suggestion for the

1354: 3-months spectrum to be harder, albeit at only \gtsima 1$\sigma$. 

1355: {\bf Bottom:} Residuals of a joint fit to the XRT spectra of J0746 with a

1356: power law and Galactic absorption only. The data from the four epochs

1357: of observations were fitted simultaneously in the energy range

1358: 2--8~keV, with the lower-energy datapoints added back. This plot

1359: clearly illustrates the spectral flattening below 3 keV at all epochs.

1360: {\it Open squares:} obsid 001; {\it Asterisks:} obsid 002; {\it Open circles:} obsid

1361: 003; {\it Filled squares:} obsid 007. }

1362: 

1363: 

1364: %\begin{figure}[h]

1365: %\begin{center}

1366: %\hbox{

1367: %% un-comment the following line to include your fig1a.ps postscript file

1368: %\psfig{figure=resid-01.ps,height=7.0cm,width=8.0cm,angle=-90}

1369: %% un-comment the following line to include your fig1b.ps postscript file

1370: %\psfig{figure=resid-02.ps,height=7.1cm,width=8.0cm,angle=-90}

1371: %}

1372: %\vspace{1.5cm}

1373: %\hbox{

1374: %\psfig{figure=resid-03.ps,height=7.1cm,width=8.0cm,angle=-90}

1375: %\psfig{figure=resid-07.ps,height=7.0cm,width=8.0cm,angle=-90}

1376: %}

1377: %\end{center}

1378: %\vspace{-1.0cm}

1379: %\caption{

1380: %{XRT spectra (top panels) and residuals (bottom panels) of the fits to

1381: %the XRT data with a single power law and fixed Galactic column

1382: %density. At all epochs, the model overestimates the soft X-ray flux,

1383: %indicating spectral flattening below 1 keV. }}

1384: 

1385: \end{figure}

1386: 

1387: \newpage 

1388: 

1389: %%%%%%%%%%%%%%%%%%%%%%%%%% figure 6: the SED %%%%%%%%%%%%%%%%%

1390: 

1391: \begin{figure}[h]

1392: \begin{center}

1393: \hbox{

1394: % un-comment the following line to include your fig1a.ps postscript file

1395: \psfig{figure=f6.ps,height=5.0in,width=5.5in}

1396: % un-comment the following line to include your fig1b.ps postscript file

1397: 

1398: %\psfig{figure=excellent_plot.ps,height=5.5cm,width=6.0cm,angle=-90}

1399: 

1400: %\vspace{1.0cm}

1401: %\psfig{figure=karen2.ps,height=4.0cm,width=4.0cm,angle=-90}

1402: }

1403: \end{center}

1404: \vspace{-1.0cm}

1405: \caption{

1406: \scriptsize

1407: {Spectral Energy Distribution of J0746 from archival and coeval

1408: observations. The BAT spectrum from the 9 months exposure is plotted,

1409: together with the XRT spectrum, UVOT, and radio (filled circles)

1410: fluxes from the November 4 campaign. The data plotted with triangles

1411: are from NED, while the GeV upper limit is from our reanalysis of the

1412: EGRET database. The SED is fitted with a composite jet+disk model. The

1413: solid line shows the jet continuum calculated with the emission model

1414: described in the text, assuming the following parameters:

1415: $\Gamma=\delta=20$, $R=2.7\times 10^{16}$ cm, $B=2$ G, electron

1416: density $n_e=9\times 10^4$ cm$^{-3}$, $\gamma _{\rm min}=1$, $\gamma

1417: _{rm b}=50$, $\gamma_{\rm max}=10^3$, $n_1=1.5$, $n_2=3.8$. The

1418: external radiation field has a luminosity $L_{\rm BLR}=10^{46}$ erg/s

1419: and is diluted within the BLR assumed to have a radius

1420: $R_{BLR}=10^{18}$ cm. The single emission components are also

1421: reportes: synchrotron (dotted), SSC (long dashed), EC (dot-dash) and

1422: the disk (dashed). The peak of the synchrotron component is produced

1423: by the self-absorption.  }}

1424: 

1425: \end{figure}

1426: \normalsize 

1427: \newpage 

1428: 

1429: 

1430: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% figure 7: SED in context %%%%%%%%%%%%%%%

1431: 

1432: 

1433: \begin{figure}[h]

1434: \begin{center}

1435: \hbox{

1436: % un-comment the following line to include your fig1a.ps postscript file

1437: \psfig{figure=f7.ps,height=14.0cm,width=15.0cm}

1438: % un-comment the following line to include your fig1b.ps postscript file

1439: %\psfig{figure=resid-02.ps,height=5.5cm,width=6.0cm,angle=-90}

1440: 

1441: %\vspace{1.0cm}

1442: %\psfig{figure=resid-03.ps,height=4.0cm,width=4.0cm,angle=-90}

1443: 

1444: }

1445: \end{center}

1446: \vspace{-1.0cm}

1447: \caption{

1448: {The Spectral Energy

1449: Distributions (SEDs) of blazars form a sequence with luminosity

1450: (Fossati et al. 1998). Going from luminous FSRQs to fainter TeV BL

1451: Lacs, the synchrotron and IC peaks move forward, and the Compton flux

1452: decreases. The solid black line is the best-fit model to the SED of

1453: J0746, showcasing its extreme properties. 

1454: }}

1455: 

1456: \end{figure}

1457: 

1458: 

1459: \end{document} 

1460: 

1461: 

1462: ******************************* OLD TEXT ***************************8

1463: Here we compare the X-ray properties of J0746 to other high-z

1464: radio-loud quasars previosuly studied at X-rays. Observations with

1465: \asca\ of a sample of such AGN showed that they have

1466: flatter X-ray continua in 0.5--10 keV than their radio-quiet

1467: counterparts, with average photon index $\Gamma_{2-10~keV}=1.53$,

1468: dispersion $\sigma_{2-10~keV}$ \ltsima 0.12 (Cappi et al. 1997). This

1469: was confirmed by recent \xmm\ observations which found

1470: $\Gamma_{2-10~keV}=1.55$ for a sample of 17 radio-loud quasars at

1471: $z>1$ (Page et al. 2005). The latter authors also found a correlation

1472: between the photon index and the luminosity, with more luminous

1473: quasars having flatter X-ray continua. At an intrinsic luminosity

1474: $(7-16) \times 10^{46}$ \lum, J0746 has a flatter 2--10 keV continuum

1475: than the radio-loud sources previously observed with \asca\ (Cappi et

1476: al. 1997) and \xmm\ (Page et al. 2005). It has been argued that the

1477: flatter X-ray spectra of radio-loud quasars compared to radio-quiet

1478: ones indicates a substantial contribution to the 2--10 keV flux by the

1479: jet (e.g., Page et al. 2005).

1480: 

1481: Similar to most $z>1$ radio-loud quasars, J0746 exhibits excess

1482: absorption along the line of sight. This seems to be a property

1483: exclusive of radio-loud sources, as no excess column densities over

1484: the galactic ones were observed in radio-quiet quasars (e.g., Fiore et

1485: al. 1998), suggesting the excess absorption is intrinsic to the

1486: source. The column density of the rest-frame absorber in J0746, N$_H^z

1487: \sim$ a few $10^{22}$ \nh, is again consistent with the values

1488: determined for other high-z sources with previous X-ray

1489: observations. While the X-ray spectrum reveals the presence of an

1490: absorber, there is no evidence for reddening in the optical spectra

1491: (Figure~1). Again, this situation is reminiscent of other AGN

1492: (REFS...). A geometrically ``stratified'' emission model, where the

1493: X-rays originate further inside the AGN than the optical has been

1494: argued as a possible viable scenario.

1495: 

1496: %However, a broken power law model fits the XRT data equally well or

1497: %better to the absorbed power law. This model yields a sharp spectral

1498: %break around 1 keV with a negative photon index below the break. In

1499: %blazars External Compton models, the latter is expected at soft

1500: %X-rays, depending very sensitively on the minimum Lorentz factor for

1501: %the emitting particles, $\gamma_{min}$, which in turn, is an essential

1502: %ingredient to calculate the jet power. 

1503: 

1504: If the medium-hard X-ray properties of J0746 are average, where this

1505: source stands out is at hard ($>$ 10 keV) energies. J0746 was one of

1506: the five brightest extragalactic sources to be detected with the BAT

1507: in the 3 months survey, with a 20--200 keV flux $\sim 10^{-10}$ \flux,

1508: correpsonding to a luminosity of $10^{48}$ \lum; its 15--150 keV

1509: spectrum is unusually flat implying that a large fraction of the

1510: quasar's bolometric luminosity during the 2005 flare was emitted above

1511: 200 keV (see below and Figure~4), most likely at GeV gamma-rays. This

1512: makes J0746 and other high-z radio-loud quasars promising candidates

1513: for detection with \glast\ (see below). 

1514: