1: \documentclass{elsart}

2: \usepackage{natbib}

3: %\usepackage{graphicx}

4: 

5: \usepackage{epsf}

6: 

7: \newcommand{\lesssim}{\lower.5ex\hbox{$\; \buildrel < \over\sim \;$}}

8: \newcommand{\gtrsim}{\lower.5ex\hbox{$\; \buildrel > \over\sim \;$}}

9: 

10: \begin{document}

11: \runauthor{Dermer and Dingus}

12: \begin{frontmatter}

13: \title{Blazar Flaring Rates Measured with {\it GLAST}}

14: \author[NRL]{C.\ D.\ Dermer,}

15: \author[LANL]{B.\ L.\ Dingus}

16: 

17: \address[NRL]{Code 7653, Naval Research Laboratory, Washington, DC 20375-5352 U.S.A.}

18: \address[LANL]{Los Alamos National Laboratory, MS H803 P-23, Los Alamos, NM 87545 U.S.A.}

19: 

20: \begin{abstract}

21: We derive the minimum observing time scales to detect a blazar at a

22: given flux level with the LAT on {\it GLAST} in the scanning and

23: pointing modes.  Based upon Phase 1 observations with EGRET, we

24: predict the {\it GLAST} detection rate of blazar flares at different

25: flux levels. With some uncertainty given the poor statistics of bright

26: blazars, we predict that a blazar flare with integral flux $\gtrsim

27: 200 \times 10^{-8}$ ph($> 100$ MeV) cm$^{-2}$ s$^{-1}$, which are the

28: best candidates for Target of Opportunity pointings and extensive

29: temporal and spectral studies, should occur every few days.

30: \end{abstract}

31: \begin{keyword}

32: Blazars, Gamma Rays

33: \end{keyword}

34: \end{frontmatter}

35: 

36: \section{Introduction}

37: 

38: The statistical significance for detection of a known source

39: at the $n\sigma$ level is $n \cong S/ \sqrt{B}$,

40: where $S$ is the number of source counts, $B$ is the number of

41: background counts, and the diffuse background is assumed to be

42: precisely known \cite{lm83}. The observer is also assumed to follow a

43: predetermined data analysis protocol and not to make multiple

44: rebinnings of the data.

45: 

46: In this paper, we derive the minimum observing time for the Large Area Telescope

47: (LAT) on {\it GLAST} (Gamma ray Large Area

48: Space Telescope) to detect

49: high-latitude $\gamma$-ray sources at given flux levels, and estimate

50: the rate at which blazar flares at certain flux levels are expected

51: to be observed with {\it GLAST}.

52: 

53: In Sections 2 and 3, we present calculations of signal and background

54: counts, respectively. The five count and the five sigma requirements

55: for detection are derived in Section 4. EGRET blazar studies,

56: described in Section 5, are used to infer the probable rate of flaring

57: emissions that GLAST will measure, as discussed in Section 6.

58: 

59: 

60: \section{Source Counts $S$}

61: 

62: The point-spread function (psf) of the LAT on {\it GLAST} is

63: characterized by angle

64: $\theta_{psf,G}=$ $(3.5^\circ/57.3^\circ)$ $(E/E_{100})^{-2/3}\equiv

65: 0.061u^{-2/3}$, where $E_{100} = 100$ MeV and $E=E$(MeV)

66: \cite{glast}. We assume an energy-dependent analysis of the {\it

67: GLAST} data, where the number of counts is integrated in the direction

68: of a catalogued source over the solid angle $\Delta\Omega_G = \pi

69: \theta_{psf,G}^2 = 0.012 u^{-4/3}$. For this analysis procedure,

70: the fraction $f_\gamma$ of detected source photons is therefore

71: independent of energy. For an azimuthally-symmetric, Gaussian

72: distribution $\propto \exp [-(\theta/\theta_{psf,G})^2]$ of count

73: directions, $f_\gamma = 1-\exp(-1) = 0.63$, though $f_\gamma\cong

74: 0.68$ would apply if $\theta_{psf}$ is defined as the 68\% containment

75: angle.

76: 

77: 

78: The number of source photon counts with energy $> E$ within $\Delta

79: \Omega_G$ of a given point source is, in this approximation,given by

80: \begin{equation}

81: S \cong f_\gamma \int _0^{\Delta t} dt \int_{E}^{\infty} 

82: dE^\prime \; A[E^\prime,\theta(t),\phi(t)]\; \phi_s(E^\prime,t)

83: \label{S}

84: \end{equation}

85: where $\phi_s(E,t)$ is the source photon flux (ph cm$^{-2}$ s$^{-1}$

86: E$^{-1}$). We approximate the energy- and angle-dependent effective

87: area of the LAT by the function

88: \begin{equation}

89: A(E) = A_0 u(\theta)(E/E_{100})^{a(\theta)}\cong X A_0 u^{a_0} \;,

90: \label{A(E)}

91: \end{equation}

92: where $u(\theta = 0^\circ) = 1$, $a(\theta = 0^\circ) = a_0$, and

93: $\theta$ is the angle between the direction of a source and the

94: axis of the LAT, and $\phi$ gives the azimuthal angle of the

95: source with respect to the LAT orientation. Here we

96: assume azimuthal symmetry of the detector effective area.  An

97: approximation for the effective area that satisfies the {\it GLAST} Science

98: Requirements Document is $A_0 \approx 6200$ cm$^2$ and $a_0

99: \approx 0.16$ for 100 MeV $\lesssim E \lesssim 10$ GeV 

100: (compare Foldout B in the {\it GLAST}

101: proposal; this approximation breaks down above 10 GeV). 

102: For 20 MeV $\lesssim E

103: \lesssim 100$ MeV, $a_0 \approx 0.37$. The net

104: observing time is denoted by $\Delta t=10^4 t_4$ s, and the exposure

105: factor X accounts for the fraction of time that a given source is

106: being detected by the LAT. The effective area of the LAT drops by a

107: factor-of-two at $\theta = 55^\circ$, so that we take $X =0.2 X_{0.2}$

108: with $X_{0.2} \cong 1$ in the scanning mode, noting that ${1\over

109: 2}(1-\cos 55^\circ )=$21\%. (Signal detection on a $10^4$ second time

110: scale is of special interest in blazar studies, because the

111: light-travel time across the Schwarzschild radius of a $10^9 M_\odot$

112: is $\simeq 10^4$ s.)

113: 

114: Normalizing the source flux to $10^{-8}\phi_{-8}~{\rm ph}(>100$ MeV)

115: cm$^{-2}$ s$^{-1}$, we have $\phi_s(E,t) = ku^{-\alpha_{ph}}$, where

116: $k = 10^{-8}\alpha \phi_{-8}/E_{100}$. For $E> 100$ MeV,

117: \begin{equation}

118: S \cong {0.15\over g}\;f_{\gamma} \;(X_{0.2}t_4) 

119: \alpha\phi_{-8} u^{\alpha_\nu-0.84}\;,

120: \label{S1}

121: \end{equation}

122: where $g = g(\alpha_\nu) = 1-(\alpha_\nu/0.84)$. Here we use the

123: following conventions to define the photon number index $\alpha_{ph}$,

124: energy index $\alpha$, and $\nu F_\nu$ index $\alpha_\nu$:

125: $\alpha=\alpha_{ph} -1$ and $\alpha_\nu = 1-\alpha = 2-\alpha_{ph}$.

126: For an intrinsic source spectrum with photon index $\alpha_{ph}\cong

127: 2$, most of the photons are detected near the lower energy threshold.

128: 

129: \section{Background Counts}

130: 

131: The number of background photons with energies $>E$ observed during

132: time $\Delta t$ that lie within solid element $\Delta \Omega$ that is

133: centered in the direction $\vec\Omega$ of a specified source is

134: \begin{equation}

135: B \cong \int_0^{\Delta t}dt\int_{E}^\infty dE^\prime \;

136: \Delta \Omega(E^\prime) \; A[E^\prime,\theta(t),

137: \phi(t)]\;\Phi_B(E^\prime,\vec\Omega)\; ,

138: \label{B}

139: \end{equation}

140: where $\Phi_B(E^\prime,\vec\Omega)$ is the diffuse $\gamma$-ray

141: background in the source direction.  A minimum level of background is

142: provided by the diffuse extragalactic $\gamma$-ray background

143: radiation, given by

144: \begin{equation}

145: \Phi_X(E) = k_x u^{-\alpha_B}\; , 0.4\lesssim u \lesssim 700

146: \label{Phi_B}

147: \end{equation}

148: \cite{sre98}, where $k_x = 1.73\pm 0.08 \times 10^{-7}$ ph

149: (cm$^2$-s-sr-MeV)$^{-1}$ and $\alpha_B = 2.10\pm 0.03$.  Other sources

150: of background, in particular, the galactic diffuse $\gamma$-ray

151: background, can also contribute; the $>100$ MeV photon flux is

152: $\approx$ twice as bright at $|b|=35^\circ$ and over 30 times brighter

153: at the Galactic center than at the poles \cite{hks97}. A better estimate

154: of blazar flaring rates must consider the galactic diffuse emission.

155: 

156: Considering only the extragalactic diffuse radiation and letting

157: $\Delta\Omega = \Delta\Omega_G$, the number of

158: background counts detected by {\it GLAST} is

159: \begin{equation}

160: B \cong 1.1 X_{0.2}t_4 u^{-2.27}\;.

161: \label{B1}

162: \end{equation}

163: This expression is valid for $1 \lesssim u \lesssim 100$.

164: 

165: \section{Five Count and 5$\sigma$ Criteria}

166: 

167: The detection of at least 5 counts implies from equation (4) that $S >

168: 5S_5$ with $S_5 \geq 1$ (the 5 count criterion), so that

169: \begin{equation}

170: (X_{0.2}t_4) \phi_{-8} \gtrsim {34\over \alpha 

171: f_\gamma}\; gS_5 u^{0.84-\alpha_\nu}\;.\;

172: \label{S5}

173: \end{equation}

174: This expression fails for hard spectrum sources with $\alpha_\nu

175: \gtrsim 0.84$, where cutoffs or breaks in the higher energy range of

176: the spectrum and high-energy corrections to the {\it GLAST} effective

177: area determine the number of detected source counts. Otherwise, most

178: of the source counts are detected at the lower energy range near $E

179: \cong E_{100}$. At energies below 100 MeV, the effective area declines

180: more rapidly, so that most photons will be detected at even lower

181: energies than 100 MeV unless $\alpha_\nu \gtrsim 0.63$.

182: 

183: Detection of a source with significance $> 5 \sigma$ requires that

184: $S/\sqrt{B} > 5\sigma_5$. Using expressions (4) and (7) for

185: $S$ and $B$, respectively, gives

186: \begin{equation}

187: (X_{0.2}t_4) \phi_{-8} \gtrsim {1220\over \phi_{-8}} 

188: \; ({g \sigma_5\over \alpha f_\gamma})^2\; u^{-2(\alpha_\nu +0.3)}\;.\;

189: \label{sigma1}

190: \end{equation}

191: By comparing equations (8) and (9), one sees that detection of

192: soft-spectrum ($\alpha_\nu \lesssim 0$) sources with the LAT at the

193: level of $\phi_{-8} \cong 15$, corresponding to the minimum flux

194: sources that EGRET could detect in a 2 week viewing period, are

195: background-limited near 100 MeV (that is, the time-on-source

196: $X_{0.2}t_4$ is larger for the $5 \sigma$ criterion when $u \approx 1$

197: than for the 5 count criterion).

198: 

199: For spectra that are not very soft (i.e., $\alpha_\nu \gtrsim -0.3$),

200: the significance of detection increases with energy even as the source

201: counting rate declines. Therefore there is a unique energy $\bar E =

202: E_{100} \bar u$ at which both the 5 count and $5 \sigma$ criteria for

203: a source emitting at a given flux level $\phi_{-8}$ are satisfied, and

204: this energy defines the minimum observing time to satisfy these

205: criteria.  For sources with $\alpha_\nu \gtrsim -0.3$, this energy is

206: \begin{equation}

207: \bar u \cong ({36 g\sigma_5^2 \over 

208: \alpha f_\gamma \phi_{-8}S_5})^{1/(1.43+\alpha_\nu)}

209: \rightarrow 16\;{\sigma_5^{1.4}\over (\phi_{-8} S_5)^{0.7}}\;,

210: \label{baru}

211: \end{equation}

212: where the last expression applies in the special case $\alpha_\nu = 0,

213: f_\gamma = 0.68$. When $\phi_{-8} = 15$, the energy to meet the two

214: criteria is $\bar u \cong 4$, or the greatest sensitivity is at

215: $\approx 200$-300 MeV.  For sources softer (more negative) than

216: $\alpha_\nu \cong -0.3$, energies near 100 MeV give the best sensitivity.

217: 

218: The minimum observing time is given by

219: \begin{equation}

220: ( X_{0.2} t_4 )_{min} = {34\over \alpha f_\gamma}\; 

221: {gS_5\over \phi_{-8}}\; \bar u^{~0.84-\alpha_\nu}\rightarrow 

222: 7 \;S_5^{0.41} \sigma_5^{1.18} ({\phi_{-8}\over 15})^{-1.59}\;,

223: \label{<Xt>}

224: \end{equation}

225: where the right-hand-side is evaluated in the special case described

226: above. Thus, {\it GLAST} will detect high-latitude sources at the

227: level EGRET could reach in two weeks in only $\sim 7\times 10^4$ s

228: $\cong 0.8$ days, and it accomplishes this task in the scanning

229: mode. Considering a background twice as large, which is more realistic

230: for the majority of blazars, gives a detection of $\phi_{-8} \cong 15$

231: sources in $\approx 1.2$ days. Consequently, {\it GLAST} will scan the

232: entire sky to the level of $\phi_{-8} \cong 15$ in a time slight

233: longer than one day. For this, we can take the EGRET phase 1 catalog

234: as a guide to what {\it GLAST} will observe.

235: 

236: \begin{figure}

237: \begin{center}

238: \vskip-1.0in

239: \leavevmode

240: \hbox{%

241: \epsfxsize=3.0in

242: \epsffile{fig1.eps}}

243: \caption{ The size distribution of blazar flares 

244: with 2-week--average fluxes $\langle \phi \rangle$ measured during the

245: phase 1 EGRET all-sky survey \cite{fic94}. The straight line

246: has slope $-3/2$.}

247: \label{fig1}

248: \end{center}

249: \end{figure}

250: 

251: \section{Comparison with EGRET Phase 1 Catalog}

252: 

253: Phase 1 of the {\it Compton Gamma Ray Observatory} observing program

254: took place during a 596 day interval from 1991 April 22 - 1992

255: November 17 \cite{fic94}.  Fig.\ 1 shows the size distribution of the

256: two-week average fluxes of radio-loud quasars and BL Lac objects

257: measured with EGRET during Phase 1, which consists of 44

258: high-confidence detections from 25 different blazars. Because the {\it

259: CGRO} Phase 1 was an all-sky survey, it will be less biased to reveal

260: the outcome of GLAST scanning observations than later EGRET catalogs

261: where the number of detections is biased by the many pointings in the

262: direction to previously known bright gamma-ray sources. Even so, the

263: Phase 1 exposure map of EGRET is far from uniform. One must keep in mind

264: that the two-week fluxes over shorter, brighter flaring phases, so that

265: the estimated blazar flaring rate could be increased if the duty cycle

266: of flares was better known cite{wal00}.

267: 

268: 

269: \begin{figure}

270: \begin{center}

271: \leavevmode

272: \hbox{%

273: \epsfxsize=3.5in

274: \epsffile{fig2.eps}}

275: \caption{Minimum required observing time for {\it GLAST} in scanning mode

276:  to satisfy 5 count detection, 5 sigma sensitivity criteria for a

277: high-latitude source at the integral photon flux level

278: $10^{-8}\phi_{-8}$ ph($>100$ MeV) cm$^{-2}$ s$^{-1}$. Right axis is

279: $\bar u = \bar E/E_{100}$, the photon energy in units of 100 MeV that

280: satisfies these criteria in the shortest time. These results are for

281: sources with flat $\nu F_\nu$ spectra, i.e., $\alpha_\nu = 0$, and assumes

282: that only the extragalactic diffuse background contributes to 

283: background counts. The solution for $\bar u$ is valid for $1 \lesssim

284: \bar u \lesssim 100$.}

285: \label{fig2}

286: \end{center}

287: \end{figure}

288: 

289: 

290: The EGRET field-of-view was about 1/24th of the full sky, so that the

291: Phase 1 duration was about twice as long as the time required to

292: survey the entire sky with two-week pointings.  During this time,

293: EGRET detected some 3-10 sources at the level $\phi_{-8} \gtrsim

294: 100$. Thus we expect {\it GLAST} to detect 2-5 such bright sources

295: during a one-day scan. Although uncertain in view of the small numbers of

296: blazars detected

297: at bright flux levels with EGRET, one

298: or two blazars brighter than $\phi_{-8} \gtrsim 200$ should be

299: detected with {\it GLAST} every 3-4 days. A comparable number of

300: unidentified $\gamma$-ray sources should also reach these levels

301: \cite{rb01}.

302: 

303: %\vskip0.1in 

304: 

305: \section{Discussion}

306: 

307: Fig.\ 2 shows the amount of time in the scanning mode required for

308: {\it GLAST} to detect sources at given flux levels, using equations

309: (10) and (11). Once the source flux $\phi_{-8} \gtrsim 200$, dozens of

310: photon counts can be expected on a $10^4$ second timescale. Changing

311: to pointing mode for a non-Earth-occulted source will increase the

312: source counts by a factor of $\sim 3$-5, with the larger number

313: applying only to those few sources that are favorably located so that

314: Earth occultation is not severe.  At these flux levels, there are now

315: sufficiently many photons to perform time-resolved

316: spectroscopy. Sources exceeding the level of $\phi_{-8}

317: \approx 200$ must be carefully and quickly reviewed to determine

318: whether to issue an alert.  Coordinated sensitive radio, optical, UV,

319: and X-ray measurements would greatly strengthen tests of dominant

320: radiation processes and inferred locations of the $\gamma$-ray

321: emission sites.  At these bright flux levels, $\gamma\gamma$

322: absorption and spectral evolution can be studied to deduce Doppler

323: factors and monitor Doppler factor variations \cite{dg95,ds02}.

324: 

325: Fig.\ 2 also shows that {\it GLAST} in the scanning mode will be

326: sensitive to high-latitude $\gamma$-ray sources at the level of

327: $\phi_{-8} = 2$ and 0.2 over $\approx 2$ weeks and 2 years observing

328: time, respectively. Thus {\it GLAST} will reach flux levels of

329: $\approx 6\times 10^{-13}$ ergs cm$^{-2}$ s$^{-1}$, fifty times

330: fainter than EGRET's limiting sensitivity. As can be seen from Fig.\

331: 2, good low-energy ($< 100$ MeV) response and calibration of the 

332: LAT on {\it GLAST} are important for identifying transients on the 

333: shortest time scales.

334: 

335: 

336: \vskip0.2in

337: 

338: We thank John Mattox for identifying an error in an early draft,

339: and acknowledge useful comments from Seth Digel.  The work of CDD is

340: supported by the Office of Naval Research and NASA {\it GLAST} science

341: investigation grant DPR \# S-15634-Y. The work of BLD is supported by

342: NASA {\it GLAST} Science Investigation grant NAG5-9712.

343: 

344: 

345: \begin{thebibliography}{99}

346: 

347: \bibitem[1]{lm83}

348: Li, P.-P., and Ma, Y.-Q. 1983, ApJ, 272, 317.

349: \bibitem[2]{glast} See LAT Instrument Performance at www-glast.stanford.edu.

350: \bibitem[3] {sre98} Sreekumar, P.~et al.\ 1998, ApJ, 494, 523 

351: \bibitem[4]{hks97}

352: Hunter, S. D., Kinzer, R. L., and Strong, A. W., 1997, in the 

353: Fourth Compton Symposium, ed. C. Dermer, M. Strickman, and J. Kurfess (AIP: New York), p. 192.

354: \bibitem[5]{fic94} Fichtel, C. E., et al. 1994, ApJS, 94, 55.

355: \bibitem[6]{wal00} Wallace, P.~M., 

356: Griffis, N.~J., Bertsch, D.~L., Hartman, R.~C., Thompson, D.~J., Kniffen, 

357: D.~A., \& Bloom, S.~D.\ 2000, ApJ, 540, 184 

358: \bibitem[7]{rb01} Reimer, O., and Bertsch, D.\ L.\ 2001, 

359:  Proc. 27th ICRC (Hamburg), 2001, pp.2566-2569 (astro-ph/0108348)

360: \bibitem[8]{dg95} Dermer, C.~D., and Gehrels, N.\ 1995, ApJ, 447, 103 

361: \bibitem[9]{ds02} Dermer, C.~D., and Schlickeiser, R.\ 2002, ApJ, 575, 667 

362: \end{thebibliography}

363: \end{document}

364: 

365: 

366: \begin{figure}[t]

367: \vspace*{2.0mm} % just in case for shifting the figure slightly down

368: \vspace*{-60.0mm}

369: %\includegraphics[width=10.0cm]{fig1.eps} % .eps for Latex,

370: %

371: \caption{The size distribution of blazar flares 

372: with 2-week--average fluxes $\langle \phi \rangle$ measured during the

373: phase 1 EGRET all-sky survey \cite{fic94}. The straight line

374: has slope $-3/2$.}

375: \end{figure}