1: %
2: %GRB predictions with GLAST
3: %
4: % Le & Dermer
5: %
6:
7: %\documentclass[10pt,preprint]{aastex}
8: \documentclass{emulateapj}
9:
10: \begin{document}
11:
12: \def\fluxthres{\hat f_{\bar \e}}
13: \def\fluxeps{f_{_{\rm \epsilon}}}
14: \def\zmin{z_{\rm min}}
15: \def\zmax{z_{\rm max}}
16: \def\xmin{x_{\rm min}}
17: \def\xmax{x_{\rm max}}
18: \def\e{\epsilon}
19: \def\Estar{{\cal E}_*}
20: \def\Estarg{{\cal E}_{*\gamma}}
21: \def\Estargo{{\cal E}_{*\gamma 0}}
22: \def\Swift{\emph{Swift}}
23:
24: \newcommand{\begeq}{\begin{equation}}
25: \newcommand{\fineq}{\end{equation}}
26: \newcommand{\begfig}{\begin{figure}}
27: \newcommand{\finfig}{\end{figure}}
28: \newcommand{\begeqarray}{\begin{eqnarray}}
29: \newcommand{\fineqarray}{\end{eqnarray}}
30:
31: \slugcomment{Accepted for publication in \apj}
32:
33: \shorttitle{GRB Predictions for the Fermi Telescope} \shortauthors{Le \& Dermer}
34:
35: \title{Gamma Ray Burst Predictions for the Fermi Gamma Ray Space Telescope}
36:
37: \author{Truong Le\altaffilmark{1} and Charles D. Dermer\altaffilmark{2}}
38: \affil{Space Science Division, Code 7653 \break Naval
39: Research Laboratory, Washington, DC 20375, USA}
40:
41: \altaffiltext{1}{tle@ssd1.nrl.navy.mil; now at Space Telescope Science Institute, Baltimore, MD}
42: \altaffiltext{2}{charles.dermer@nrl.navy.mil}
43:
44: \begin{abstract}
45: Results of a phenomenological model to estimate the GRB detection
46: rate by the Fermi Gamma ray Space Telescope are reported. This
47: estimate is based on the BATSE 4B GRB fluence distribution, the
48: mean ratio of fluences measured at $100$ MeV -- 5 GeV with EGRET
49: and at $20$ keV -- 2 MeV with BATSE, and the mean EGRET GRB
50: spectrum for the 5 EGRET spark-chamber GRBs. For a 10\% fluence
51: ratio and a number spectral index $\alpha_1 = -2$ at 100 MeV -- 5
52: GeV energies, we estimate a rate of $\approx 20$ and 4 GRBs per yr
53: in the Fermi Large Area Telescope field of view with at least 5
54: photons with energy $E > 100$ MeV and $E > 1$ GeV, respectively.
55: We also estimate $\approx 1.5$ GRBs per yr in the Fermi FoV where
56: at least 1 photon with energy $E > 10$ GeV is detected. For these
57: parameters, we estimate $\approx 1$ -- 2 GRBs per year detected
58: with the Fermi telescope with more than 100 $\gamma$ rays with
59: $E\gtrsim 100$ MeV. Comparison predictions for $\alpha_1 = -2.2$,
60: different fluence ratios, and the AGILE $\gamma$-ray satellite are
61: made. Searches for different classes of GRBs using a diagram
62: plotting $100$ MeV -- 10 GeV fluence vs.\ 20 keV -- 20 MeV fluence
63: is considered as a way to search for separate classes of GRBs and,
64: specifically, spectral differences between the short-hard and long
65: duration GRB classes, and for hard components in GRBs.
66: \end{abstract}
67:
68: \keywords{gamma-rays: bursts---theory }
69:
70: \section{Introduction}
71:
72: The Gamma ray Large Area Space Telescope, GLAST, was launched on
73: 11 June 2008 and renamed the Fermi Gamma ray Space Telescope
74: (hereafter Fermi) on August 26, 2008. Fermi comprises two
75: separate instruments, the Large Area Telescope (LAT), which is
76: sensitive to $\gamma$ rays in the energy range from $E\approx 30$
77: MeV to $ E >$ 300 GeV, and the Fermi GLAST Burst Monitor (GBM),
78: sensitive at 10 keV $\lesssim E \lesssim$ 30 MeV. At GeV energies,
79: the Fermi LAT provides an increase in effective area over the
80: Energetic Gamma Ray Experiment Telescope (EGRET) on the Compton
81: Gamma Ray Observatory (CGRO) by almost an order of magnitude, a
82: smaller point spread function, a field-of-view (FOV) larger by a
83: factor $\approx 5$ than EGRET, and the capability to autonomously
84: slew the LAT in response to a trigger from the GBM. The GBM uses
85: NaI scintillators for triggering, like the Burst and Transient
86: Source Experiment (BATSE), in addition to BGO (Bismuth Germanate)
87: detectors for response in the $\sim 0.2$ MeV -- 25 MeV energy
88: range. The GBM views the unocculted sky, $\approx 50$\% of the
89: full sky, and the predicted long-duration GBM burst detection rate
90: is $\approx 150$ -– 225 per year when triggering using standard
91: BATSE detection criteria \citep[$>4.5\sigma $ in at least two
92: detectors in 1.024 s in the 50 –- 300 keV;][]{kip01}. This can be
93: compared with the BATSE detection rate of $\approx 550$ GRBs
94: full-sky brighter than $0.3$ cm$^{-2}$ s$^{-1}$ in the 50 -- 300
95: keV band \citep{ban02}.
96:
97: Five EGRET $\gamma$-ray transients, coincident in time and
98: direction with a BATSE GRB, were detected with the EGRET spark
99: chamber at $E> 30$ MeV with photon counts in excess of background
100: \citep{din95}. Knowledge from the CGRO can be used to quantify
101: expectations for Fermi. To simulate GRBs in the Fermi energy band,
102: we calculate the average fluence ratio between EGRET (100 MeV -- 5
103: GeV) and BATSE ($\simeq 20$ keV -- $2$ MeV) from the 5 BATSE GRBs
104: that were detected with the EGRET spark chamber. Using an average
105: spectrum consisting of an empirical \citet{ban93} spectrum plus a
106: power law, and the effective area of the Fermi LAT, we estimate
107: the full-sky number of GRBs per year with different numbers of
108: high-energy photons based on the BATSE 4B fluence distribution for
109: long-duration GRBs \citep{pac99}. We make predictions for both
110: long-duration and total GRBs; only the former class of GRBs were
111: convincingly detected at 100 MeV -- 5 GeV energies with CGRO. Both
112: Fermi and AGILE view $\approx 1/5^{th}$ of the full sky at 100 MeV
113: -- 5 GeV energies, with the field-of-view defined by the solid
114: angle where the effective area is $\gtrsim$ one-half the on-axis
115: effective area. Thus we consider the realistic detection rate to
116: be $\approx 0.2\times$ the full-sky rate, without considering the
117: potential effects of autonomous repointings in the estimates.
118:
119: In \S~2 we discuss the phenomenological model for GRBs based on
120: BATSE and EGRET data, and estimate the uncertainty in GRB
121: properties derived from EGRET and Fermi measurements. The
122: predicted GRB detection rates and GRB contribution to the diffuse
123: extragalactic background radiation are calculated in \S~3. We note
124: that these predictions were made in advance of the launch of Fermi
125: (see arXiv:0807.0355 v1), so can be used to test our knowledge of
126: GRB physics, in particular, whether additional hard, high-energy
127: ($\gtrsim 30$ MeV) $\gamma$-ray spectral components are present in
128: GRBs.
129:
130: \section{Model}
131:
132: We develop a phenomenological model of GRBs based on the
133: properties of EGRET spark-chamber GRBs, all detected with BATSE,
134: and then make predictions for the GRB detection rate with Fermi
135: and AGILE. Based on BATSE and EGRET results, and from earlier GRB
136: studies, the time-integrated (and resolved, where possible) GRB
137: photon number spectrum $N(E_\gamma )$ is assumed to be
138: well-described by the sum of the Band function \citep{ban93},
139: $N_{\rm B}(\e )$, and a high-energy power law component,
140: $N_G(\epsilon )$. We write this function using dimensionless
141: notation $\epsilon = E_\gamma / m_e c^2$ as
142: %
143: \begin{equation}
144: N(\epsilon ) = N_{\rm B}(\epsilon ) + N_G(\epsilon )
145: \label{nepsilon}
146: \end{equation}
147: %
148: where $N_{\rm B}(\e)$ takes the form
149: %
150: \begin{eqnarray}
151: N_{\rm B}(\epsilon) & = & k_{_B} \; \epsilon^{\alpha}
152: \exp[-\epsilon (\alpha-\beta)/\epsilon_{_{\rm br}}] \; H(\epsilon;
153: \e^{\rm B}_{\rm min},\epsilon_{_{\rm br}}) \nonumber \\
154: & + & \; k_{_B} \; \epsilon^\beta \; \epsilon_{_{\rm br}}^{\alpha-\beta}
155: \exp(\beta-\alpha)
156: \; H(\e;\e_{\rm br},\e^{\rm B}_{\rm max}) \; . %
157: \label{Nband-eq} %
158: \end{eqnarray}
159: %
160: The parameters $\alpha$ and $\beta$ are the low- and high-energy
161: Band $\alpha$ and Band $\beta$ indices, and $E_{\rm br} =
162: m_ec^2\epsilon_{_{\rm br}}\sim 100$ keV is the ``break energy.''
163: The photon energy $E_{\rm pk}$ of the peak of the $\nu F_\nu$
164: spectrum is related to $E_{\rm br}$ through the expression
165: %
166: \begin{equation}
167: E_{\rm pk} = {(2+\alpha)E_{\rm br}\over \alpha - \beta }\;,
168: \label{Epk}
169: \end{equation}
170: %
171: provided $\alpha >-2$ and $\beta <-2$. The Heaviside function is
172: defined such that $H(x;y,z)=0$ except at $y\leq x\leq z$, where
173: $H(x;y,z)=1$ . The term $k_{_B}$ is the constant normalizing the
174: number fluence to the $> 20$ keV BATSE energy fluence $\Phi_{_B}
175: (> 20 {\rm \; keV})$ of a particular GRB, and is given by $k_{_B}
176: = \Phi_{_B} (> 20 {\rm \; keV})/I_{_1}$, where
177: %$$$$
178: \begin{eqnarray}
179: {I_{_1}\over m_e c^2} & = & \int^{\epsilon_{_{\rm
180: br}}}_{\epsilon^{\rm B}_{_{\rm min}}} \epsilon^{\alpha + 1}
181: e^{-\epsilon (\alpha - \beta)/\epsilon_{_{\rm br}}} \; d\epsilon\ \nonumber \\
182: & + & \; \epsilon_{_{\rm br}}^{\alpha - \beta} e^{\beta - \alpha}
183: \int^{\epsilon^{\rm B}_{_{\rm max}}}_{\epsilon_{_{\rm
184: br}}} \epsilon^{\beta + 1} \; d\epsilon \; . %
185: \label{I1-eq} %
186: \end{eqnarray}
187:
188: For model GRBs normalized to the BATSE fluence, we take the
189: minimum energy $\epsilon^{\rm B}_{_{\rm min}}=30$ keV, and maximum
190: energy $\epsilon^{\rm B}_{_{\rm max}}= 5$ GeV. For the mean GRB
191: spectral indices, we take $\alpha = -0.9$, $\beta = -2.2$, and
192: break energy $E_{_{\rm br}}= 250$ keV, corresponding to the mean
193: for these values from spectral analyses of bright BATSE GRBs
194: \citep{pre00}.
195:
196: For the high-energy spectral component, we assume that GRBs can be
197: described by a power law at $\gamma$-ray energies written as
198: %
199: \begin{equation}
200: N_{_G}(\epsilon) = k_{_G} \epsilon^{\alpha_{_1}} H(\epsilon; \;
201: \epsilon_{_{\rm min}}, \epsilon_{_{\rm max}}) \; .%
202: \label{NG-eq} %
203: \end{equation}
204: %
205: The average spectrum of GRBs measured by the EGRET spark chamber
206: is $\alpha_1 = -1.95\pm 0.25$ for 45 $\gamma$ rays detected within
207: 200 seconds of the trigger \citep{din95,din98}, similar to the
208: average EGRET Total Absorption Shower Counter (TASC) index
209: \citep{cds98}. Band $\beta$ values for BATSE GRBs have a large
210: range, with most GRBs showing $ -2 \lesssim \beta \lesssim 2.5$,
211: but lie in a different range than EGRET and Fermi emissions. The
212: mean spectral index measured by EGRET for the five spark-chamber
213: GRBs is not in good agreement with the index of BATSE and TASC
214: GRBs at $E\gtrsim 1$ MeV \citep{pre00}. This suggests that the
215: spectrum is probably not a simple power law at high energies, but
216: different spectral shapes should be considered \citep[e.g.,
217: cut-offs or extra components; see][]{kgp08}, and unusual spectral
218: evolution of high-energy components \citep{gon03}.
219:
220: Since the spectral index distribution of GRBs at medium
221: $\gamma$-ray energies is still debatable, which Fermi will
222: clarify, in this paper we make predictions assuming that the
223: $>100$ MeV $\gamma$-ray spectrum can be described by a simple
224: power law with $\alpha_1 \cong -2$. For definiteness, we consider
225: $\alpha_1 = -2$ and $\alpha_1 = -2.2$. The latter value
226: corresponds to the softest spectrum within 1$\sigma$ of the EGRET
227: measurements, and is the same as the Band $\beta$ used,
228: effectively giving a single power law without break. Besides
229: spectral index, our predictions are also strongly affected by
230: EGRET uncertainties, which we will discuss in the next section.
231:
232: The term
233: \begin{equation}
234: k_{_G}= {\rho \Phi_{_B}(> 20
235: {\rm~keV})\over I_{2}}\;
236: \label{kG}
237: \end{equation}
238: where
239: \begin{equation}
240: {I_{2}\over m_e c^2} = \int^{\epsilon{_{\rm max}}}_{\epsilon_{_{\rm
241: min}}} \;d\e \;\epsilon^{\alpha_{_1} + 1}
242: \label{I_2}
243: \end{equation}
244: is the normalization constant based on the average fluence ratio
245: $\rho$ between EGRET and BATSE. In our model, the uncertainties in
246: the number of $\gamma$ rays that were detected by EGRET and the
247: measured BATSE and EGRET fluences are absorbed into this factor
248: $\rho$. Here $\epsilon_{_{\rm min}}$ and $\epsilon_{_{\rm max}}$
249: are the minimum and maximum energy at 100 MeV and 5 GeV,
250: respectively; note that even though EGRET detected photons down to
251: 30 MeV energies, the effective area dropped rapidly below $\approx
252: 100$ MeV. The higher energy given by $\epsilon_{_{\rm max}}$ is
253: set by the energy where self-vetoing in EGRET reduces its
254: effective area (see Fig.~\ref{fig1}).
255:
256: %
257: \begin{figure}[t]
258: \includegraphics[width=3.0in]{f1.eps}
259: \caption{Approximate on-axis effective area of the Fermi LAT, EGRET, and AGILE.
260: The symbols are the data, and
261: the solid lines are the best-fitted functions.\\} %
262: \label{fig1} %
263: \end{figure}
264: %
265:
266: The number of source counts with energy $\epsilon >
267: \epsilon_{\gamma}$ for different detectors \citep{dd04} is given
268: by
269: %
270: \begin{equation}
271: S_i(>\epsilon_\gamma) \cong f_{_\gamma}
272: \int^{\infty}_{\epsilon_{\gamma}} A_i(\epsilon) \;
273: N(\epsilon) \; d\epsilon \; , %
274: \label{Source-eq} %
275: \end{equation}
276: %
277: where $A_i(\e )$ is the on-axis effective area of different
278: detectors (e.g., EGRET, Fermi LAT, and AGILE) and $f_{_\gamma}$ is
279: a collection function depending on the opening angle of the
280: detector. Here we conservatively take the opening angle as
281: energy-dependent with $f_{_\gamma} = 68\%$ containment
282: \citep{tho86,dd04} though a larger region of interest (ROI) will
283: not much change the results for bright GRBs with many $\gamma$
284: rays. The on-axis effective areas for EGRET ($A_E$), Fermi LAT
285: ($A_F$), and AGILE ($A_A$) are taken from the best fit to the
286: calibration and analysis data \citep[][and references
287: therein]{pit04,bel07} as shown in Fig.\ \ref{fig1}. The functions
288: describing these areas are
289: %
290: \begin{eqnarray}
291: A_E(x) & = & 10^{a_0 + a_1 x + a_2 x^2 + a_3 x^3 + a_4 x^4} \nonumber \\ %
292: A_F(x) & = & 10^{a_0 + a_1 x + a_2 x^2 + a_3 x^3 + a_4 x^4} \\
293: A_A(x) & = & 10^{a_0 + a_1 x + a_2 x^2} \; , \nonumber %
294: \label{Earea-eq} %
295: \end{eqnarray}
296: %
297: where $x=\log E$, and parameters are given in Table~1. Most GRBs
298: would take place at relatively large angles, $\approx 30$ --
299: 60$^\circ$, from the Fermi axis, where the effective area is half
300: this value. On the other hand, a larger ROI would include more
301: counts. These represent two additional uncertainties in the rate
302: estimates.
303:
304: \begin{deluxetable}{cccccc}
305: \tabletypesize{\small} %
306: \tablecaption{Effective Area Parameters\label{tbl-1}} %
307: \tablewidth{-0.pt} %
308: \tablehead{ %
309: \colhead{$\rm Effective \atop {Area \atop (cm^2)}$} %
310: &\colhead{$a_0$} %
311: &\colhead{$a_1$} %
312: &\colhead{$a_2$} %
313: &\colhead{$a_3$} %
314: &\colhead{$a_4$} %
315: } \startdata
316: $A_E$ %
317: & $-24.116 $ %
318: & $36.435 $ %
319: & $-18.000 $ %
320: & $3.912 $ %
321: & $-0.317 $ %
322: \\
323: $A_A$ %
324: & $2.010 $ %
325: & $0.501 $ %
326: & $-0.080 $ %
327: \\
328: $A_F$ %
329: & $-11.583 $ %
330: & $18.503 $ %
331: & $-8.127 $ %
332: & $1.555 $ %
333: & $-0.109 $ %
334: \enddata
335: \end{deluxetable}
336:
337: The on-axis effective area of Fermi is shown in Fig.\ \ref{fig1}.
338: Signal analysis of photon events in the Fermi LAT depends on event
339: reconstruction and background rejection \citep{atw09}. The {\it
340: diffuse} class of events, for instance, minimizes background by
341: rejecting tracks without good energy measurement in the
342: calorimeter and requires both tracker and calibration information.
343: GRBs can, however, be studied with the {\it transient} event
344: class, which increases low-energy response at the expense of
345: greater background using large regions of interest around the GRB
346: direction. We consider only photons with $E > 100$ MeV, where the
347: point spread function and consequently the background is smaller
348: than for photons at lower energies, limited at higher energies to
349: a maximum of 5 GeV. Over this energy range, both the EGRET and
350: Fermi effective area is relatively flat, and measurement
351: uncertainty can be more reliably estimated.
352:
353: \begin{deluxetable*}{lccccccccccc}
354: \hskip-1.in
355: \tabletypesize{\scriptsize} %
356: \tablecaption{EGRET Gamma-Ray Burst Spark Chamber Detections\label{tbl-2}} %
357: \tablewidth{0pt} %
358: \tablehead{ %
359: \colhead{} & %
360: \colhead{} & %
361: \colhead{} & %
362: \colhead{} &%
363: \colhead{} & %
364: \colhead{} & %
365: \multicolumn{2}{c}{EGRET GRB $\gamma$ rays} & %
366: \multicolumn{2}{c}{spectral index of $-2$} & %
367: \multicolumn{2}{c}{spectral index of $-2.2$} \\ %
368: \colhead{$\rm Burst \atop Date$} & %
369: \colhead{$\theta^a$} & %
370: \colhead{$N_{sc}^b$} & %
371: \colhead{$N_{>100}^c$} & %
372: \colhead{$E_{max}({\rm GeV})$} & %
373: \colhead{$\Phi^d_{_{\rm BATSE}} $} & %
374: \colhead{$\Phi^e_{_{\rm EGRET}} $} & %
375: \colhead{$\rho(\% )^f$} & %
376: \colhead{$\Phi^e_{_{\rm EGRET}} $} & %
377: \colhead{$\rho(\% )^f$} & %
378: \colhead{$\Phi^e_{_{\rm EGRET}} $} & %
379: \colhead{$\rho(\%)^f$} %
380: } %
381: \startdata
382: 910503 %
383: & 24 %
384: & 9$^g$ %
385: & 2 %
386: & 10 %
387: & $2.92(\pm 0.01) $ %
388: & $1.28$ %
389: & $0.43$ %
390: & $2.68$ %
391: & $0.92$ %
392: & $2.20$ %
393: & $0.75$
394: \\
395: 910601 %
396: & 12 %
397: & 8 %
398: & 3 %
399: & 0.31 %
400: & $1.5 (\pm 0.01)$ %
401: & $1.13$ %
402: & $0.75$ %
403: & $1.79$ %
404: & $1.19$ %
405: & $1.48$ %
406: & $0.99$
407: \\
408: 930131 %
409: & 28 %
410: & 18 %
411: & 12 %
412: & 1.2 %
413: & $0.66(\pm 0.11)$ %
414: & $26.6$ %
415: & $40.3$ %
416: & $24.8$ %
417: & $37.6$ %
418: & $20.3$ %
419: & $30.8$
420: \\
421: 940217 %
422: & 10 %
423: & 28 %
424: & 10 %
425: & 18 %
426: & $6.6 (\pm 0.03) $ %
427: & $13.0$%
428: & $1.97$ %
429: & $5.53$ %
430: & $0.84$ %
431: & $4.58$ %
432: & $0.69$
433: \\
434: 940301 %
435: & 6 %
436: & 7 %
437: & 5 %
438: & 0.16 %
439: & $1.12(\pm 0.01) $ %
440: & $1.14$ %
441: & $1.02$ %
442: & $2.48$ %
443: & $2.21$ %
444: & $2.05$ %
445: & $1.83$
446: \enddata
447: %% Text for table notes should follow after the \enddata but before
448: %% the \end{deluxetable}. Make sure there is at least one \tablenotemark
449: %% in the table for each \tablenotetext.
450: \tablecomments{
451: \noindent $^a$Angle from axis in degrees\\
452: \noindent $^b$Number of $\gamma$ rays imaged by EGRET spark chamber\\
453: \noindent $^c$Number of $\gamma$ rays imaged by EGRET spark
454: chamber between 100 MeV and 5 GeV \\
455: \noindent $^d$BATSE 20 keV -- 2 MeV fluence, units of $10^{-4}$
456: erg cm$^{-2}$, from \citet{mee96}\\
457: \noindent $^e$EGRET 100 MeV -- 5 GeV fluence, units of $10^{-6}$ erg cm$^{-2}$\\
458: \noindent $^f\rho={\Phi_{\rm EGRET}}/{\Phi_{\rm BATSE}} $\\
459: \noindent $^g$Nine events within 15 s after GRB trigger
460: \citep{sch92}}
461: \end{deluxetable*}% table 3 follows
462: %
463:
464: \section{Fluence Estimates and Uncertainties for EGRET GRBs}
465:
466: In Table~2, we list the pertinent data from EGRET to make
467: estimates of Fermi detection rates. The first five columns of
468: Table~2 give the GRB date-name, the angle of the GRB with respect
469: to the z-axis of the EGRET telescope, the total number and the
470: number of $\gamma$ rays with energies between 100 MeV and 5 GeV
471: imaged by the EGRET spark chamber, and the maximum $\gamma$-ray
472: energy associated with the GRB. With this data, we can calculate
473: the 100 MeV -- 5 GeV fluence $F_{0.1{\rm-}5{\rm~ GeV}}$ from an
474: EGRET spark chamber GRB using the expression
475: %
476: \begin{equation}
477: F_{0.1-5{\rm~ GeV}}= \sum_i{E_i\over A(E_i,\theta)}\;,
478: \label{fluence}
479: \end{equation}
480: %
481: with the sum being over the $i = 1,\dots, N_\gamma$ $\gamma$ rays
482: with energies from 100 MeV to 5 GeV detected by the EGRET spark
483: chamber and associated with the GRB, located at the angle $\theta$
484: with respect to the pointing axis of the EGRET.
485:
486: By choosing a specific energy window between 100 MeV and 5 GeV, we
487: can minimize uncertainty in the GRB fluence estimate. This is
488: where the EGRET response is best understood, as compared to the
489: higher energy regime with uncertain systematics \citep{shk08}, or
490: at lower energies where the point spread function increases and
491: the numbers of background photons are larger. Because the
492: background photons are more likely to be at low energies, for
493: example, due to difficulties in event reconstruction, the number
494: of background photons, quoted below from the analysis of each
495: specific EGRET GRB, is an upper limit to the number of background
496: photons in the 100 MeV -- 5 GeV range.
497:
498: \subsection{Fluence Error Estimate for EGRET Spark Chamber GRBs}
499:
500: Several errors come into the estimate of the 100 MeV -- 5 GeV
501: burst fluence, including (i) the statistical error due to the
502: finite number of detected photons; (ii) deadtime effects in the
503: EGRET detector; and (iii) background spark-chamber events
504: erroneously associated with the GRB. Other effects may also play a
505: role, including (iv) errors in energy resolution ($\approx 12$ \%
506: and 9\% $\gamma$-ray photon energy resolution errors at 100 MeV
507: and 1 GeV, respectively), and (v) errors in effective area. For
508: the EGRET effective area, we use the angle-dependent EGRET
509: effective area with the TASC in coincidence \citep[Fig.\ 14
510: of][]{tho93}, where the effective area for the associate burst
511: (See Table~\ref{tbl-2}) is obtained by interpolating the data in
512: Figure~14.
513:
514: For GRB 910503, 9 photons are detected within 15 s of the GRB
515: trigger, and 6 are likely to be associated with this GRB, implying
516: a possible background of 3 events \citep{sch92}. A 10 GeV $\gamma$
517: ray was also detected 84 s after this GRB \citep{mer95} coincident
518: with its location. Six of the $\gamma$ rays were detected within 2
519: seconds, allowing as much as a factor of $\approx 30$\% error for
520: 0.1 s EGRET deadtime if high-energy $\gamma$ rays are produced
521: roughly uniformly over this time interval. If such emission is
522: produced in bursts shorter than the $\sim 0.1$ s deadtime, then
523: the GRB produced a stronger fluence then measured. For GRB 910601
524: \citep{kwo93}, 4 events were detected within 100 s of the GRB
525: trigger, with 3 events greater $100$ MeV $\gamma$ rays, when only
526: 1.5 events were expected. This represents a significant
527: uncertainty on the fluence due to background. In contrast,
528: deadtime effects are unlikely to be important for GRB910601
529: assuming, as we do henceforth, that $\gamma$-rays are not produced
530: in sub-100 ms bursts.
531:
532: GRB 930131, the so-called ``superbowl burst," is interesting in
533: that it shows an intense $\sim 0.1$ s peak characteristic of a
534: short hard GRB \citep{som94}, followed $\approx 0.8$ s later by a
535: weaker $\approx 0.4$ s pulse, and also displays extended emission
536: detected with BATE for ~50 s after the GRB trigger \citep{kou94}.
537: Although its $t_{90}$(50 -- 300 keV) $= 19.200\pm 2.56$ s, its
538: $t_{50}$(50 -- 300 keV) $ = 1.024\pm 0.091$ s \citep{pac99}, so it
539: is near the boundary between the long soft and short hard GRBs
540: \citep{kou93}. Note that short hard GRBs are often observed with
541: Swift to produce extended emission \citep{ng08}. Sixteen $\gamma$
542: rays are imaged within 25 s of the BATSE trigger, of which two
543: pairs and a cluster of three arrive on a sufficiently short,
544: $\lesssim 0.2$ s timescale that deadtime effects could have
545: reduced the detected number of photons. Based on background
546: observed before the trigger time of the GRB, only 0.04 events are
547: expected by chance within 25 s \citep{som94}, so background should
548: not be important for this GRB. This GRB was very intense at BATSE
549: energies, but because most of this emission appeared in a first
550: very short, very intense pulse, the BATSE 20 keV -- 2 MeV fluence
551: is smallest of the 5 spark-chamber GRBs.
552:
553: GRB 940217 \citep{hur94}, with spark-chamber emission extending to
554: $> 100$ min after the GRB trigger, has associated with it the
555: highest energy photon of all EGRET GRBs, $\approx 18$ GeV. A total
556: of 28 $\gamma$-ray photons were imaged by the EGRET spark chamber
557: from this GRB, with 10, 8, and 10 appearing during the $\approx
558: 90$ s episode of strong BATSE emission, after the BATSE emission
559: and before Earth occultation occurring at $\approx 900$ s after
560: the GRB trigger, and after the GRB appears from Earth occultation
561: at $\sim 3700$ to 5400 s after the GRB. For these three periods,
562: respectively, 0.39, 1.8, and 2.9 background $\gamma$ rays are
563: expected. Because the GRB was so extended in time and the $\gamma$
564: rays arrive one or more seconds apart, deadtime effects can be
565: considered to be small for this GRB. The duration of Earth
566: occultation, representing $\gtrsim 62$\% of the known period of
567: activity of GRB 940217, therefore represents a major uncertainty
568: in the fluence measurement of this GRB.
569:
570: Seven photons were imaged within 21 s of the GRB 940301 BATSE
571: trigger \citep{sch95}, though appearing sparsely distributed in
572: time so that deadtime effects are probably negligible. Background
573: is also likely to be small for this GRB, as no other $\gamma$ rays
574: are detected over a period of 150 s around the time of the GRB
575: except for the 7 photons mentioned.
576:
577: For completeness, two other GRBs from the CGRO era should be
578: mentioned. This includes GRB 910814, with $\gamma$ rays observed
579: up to $\approx 50$ MeV with the TASC \citep[when the spark chamber
580: was disabled due to Earth occultation; in any case this event was
581: at 38.7$^\circ$ from the EGRET boresight;][]{kwo93}. The $>20$ keV
582: fluence of GRB 910814 was $\cong 2.51\times 10^{-4}$ ergs
583: cm$^{-2}$ \citep{mee96}. Late in the CGRO mission, when the EGRET
584: spark chamber gas was mostly depleted, GRB 990123 was observed
585: with COMPTEL to $\approx 18$ MeV. This was a bright GRB, with a
586: $>20$ keV fluence $\cong 5.1\times 10^{-4}$ ergs cm$^{-2}$
587: \citep{bri99}. At redshift $z = 1.6$ \citep{kul99}, it provided
588: the first confident lower limits on bulk Lorentz factor using
589: internal $\gamma\gamma$ opacity arguments \citep{bar06}. Joint
590: analysis of BATSE and EGRET-TASC GRBs indicates excess, $\gtrsim
591: 1$ MeV emission, though not necessarily extending to $\gtrsim 100$
592: MeV, in bright, long duration GRBs \citep{gon09}. The $>20$ keV
593: fluence of the most notable of these bursts, GRB 941017, was
594: $\cong 3\times 10^{-4}$ ergs cm$^{-2}$.
595:
596: %
597: \begin{figure}[t]
598: \includegraphics[width=3.1in]{f2.eps}
599: \caption{Histogram of BATSE 4B catalog GRB fluence distribution;
600: data from \citet{pac99}. %
601: \label{fig2}} %
602: \end{figure}
603: %
604:
605: Comparison of the distribution of the fluences of GRBs shown in
606: Fig.\ \ref{fig2} in the BATSE 4B catalog with the fluences of the
607: EGRET spark chamber GRBs in Table 2, and notable BATSE-TASC GRBs
608: in the preceding paragraph, shows that the high-energy emission
609: is, for the long duration GRBs, always associated with the most
610: fluent GRBs. This is characteristic of a sensitivity limited
611: detection of high-energy $\gamma$-ray emission. Consequently, we
612: can use the increased capabilities of the Fermi telescope with
613: respect to EGRET to make straightforward predictions by scaling
614: from the EGRET results. Note that the weakest GRB in Table 2 in
615: terms of fluence---GRB 930131---is also possibly a short, hard
616: GRB.
617:
618: \subsection{Error from Finite Photon Number}
619:
620: We consider a major error on fluence measurement to result from
621: statistical fluctuations due to the small number of photons imaged
622: by the EGRET spark chamber and associated with the GRB,
623: particularly since we limited the spark chamber events to those
624: with energy between 100 MeV and 5 GeV. It is a simple matter to
625: make this error estimate using a Monte Carlo simulation, and then
626: add other errors to get a complete error estimate.
627:
628: Let the integrated GRB photon number spectrum be denoted by
629: $dN/dEdA$. The fluence measured between energies $E_1$ and $E_2$
630: is given by
631: %
632: \begin{equation}
633: F(E_1,E_2) = \int_{E_1}^{E_2} dE \; E \left({dN\over dE dA}\right)\;.
634: \label{FE1E2}
635: \end{equation}
636: %
637: The number of source photons is given by
638: %
639: \begin{equation}
640: N_s(E_1,E_2) = \int_{E_1}^{E_2} dE \; A(E,\theta) \left({dN\over dE dA}\right)\;.
641: \label{NsE1E2}
642: \end{equation}
643: %
644: The probability of detecting a photon of energy $E$ is simply
645: given by inverting the expression
646: %
647: \begin{equation}
648: r = {\int_{E}^{E_2} dE \; A(E,\theta) (dN/dE dA)\over \int_{E_1}^{E_2} dE \; A(E,\theta) (dN/dE dA)}\;,
649: \label{rinvert}
650: \end{equation}
651: %
652: for $E$, where $r$ is a random number uniformly distributed
653: between 0 and 1.
654:
655: In the EGRET (and Fermi) energy range, $A(E,\theta)\cong const$
656: within a factor of $\approx 1.8$ for $100 < E({\rm MeV}) < 5000$
657: and $\theta < 30^\circ$, and can be approximated by a power law
658: over this range, so we write $A(E,\theta) = A_0
659: E^{-u}$.\footnote{See Fig.\ \ref{fig1} and
660: www-glast.slac.stanford.edu/software/IS/glast\_\\
661: lat\_performance.htm; restricting the photon energies to $>200 $
662: MeV makes this an excellent approximation.} The effect of changing
663: $\theta$ on $A(E,\theta)$ over the duration of the GRB is
664: negligible for EGRET, which was a pointing telescope, but can be
665: important in Fermi observations of GRBs. More accurate numerical
666: calculations can be made following the analytic inversion given
667: here.
668:
669: Assuming that $dN/dEdA \propto E^{-\alpha_1}$, the inversion of
670: eq.\ (\ref{rinvert}) is trivial, and we obtain for the detected
671: energy the value
672: %
673: \begin{equation}
674: E(r) = \left[ E_2^{1-s} +r(E_1^{1-s} - E_2^{1-s})\right]^{1/(1-s)}\;, %
675: \label{E(r)}
676: \end{equation}
677: %
678: where $s = u +\alpha_1$. The mean photon energy in the energy
679: range $E_1 \leq E \leq E_2$ is
680: %
681: \begin{equation}
682: \langle E\rangle = \left( {E_1^{2-\alpha_1} - E_2^{2-\alpha_1}\over {E_1^{1-\alpha_1} - E_2^{1-\alpha_1}} }\right)
683: \;\left({\alpha_1 - 1\over \alpha_1-2}\right)\;.
684: \label{meanenergy}
685: \end{equation}
686: %
687: The error in fluence due to the finite number, $N_\gamma$, of
688: photons detected is then given by the root mean square deviation
689: of the detected photon energies from the average photon energy,
690: that is,
691: \begin{equation}
692: \Delta E = \sqrt{ { \sum_{i = 1}^{N_\gamma}[E(r_i) - \langle E\rangle]^2\over N_\gamma} }\;.
693: \label{fluence_error}
694: \end{equation}
695:
696: \begin{figure}[t]
697: \includegraphics[width=3.50in]{f3.eps}
698: \caption{Fractional uncertainty in measurement of the 100 MeV -- 5
699: GeV $\gamma$-ray fluence due to finite number of photons from the
700: GRB, for an underlying photon spectra described as a power-law
701: with index $\alpha_1 = p$. Calculation uses $N = 1000$ sets of
702: $N_\gamma$ photons chosen randomly, except where noted.%
703: \label{fig3new}}
704: \end{figure}
705:
706: The fractional fluence error, $\Delta F/F$, as a function of the
707: number of detected photons $N_\gamma(100$ MeV -- 5 GeV) in the
708: photon energy range between $E_1 = 100$ MeV to $E_2 = 5$ GeV range
709: is shown in Fig.\ \ref{fig3new} for the simplest approximation of
710: a flat effective area, so that $u = 0$. Because the fluence is
711: weighted by energy, and we have restricted the photon energy to a
712: finite range, a spectrum with $\alpha_1 = 2$ gives the largest
713: fluence uncertainty, which is well described by the
714: function
715:
716: %
717: \begin{equation}
718: {\Delta F\over F} \cong {1.45\over \sqrt{N_\gamma}}\;,
719: \label{dFoverF}
720: \end{equation}
721: %
722: where that the factor of 1.45 is derived from fitting the data
723: (with p = 2) in Figure~\ref{fig3new}. We can take this expression
724: as providing the maximum error in fluence from the statistical
725: uncertainty due to the finite number of photons. If there is a
726: priori knowledge that the intrinsic spectrum has a specific number
727: index, e.g., $\alpha_1 = 2.2$, then the numerical coefficient of
728: eq.\ (17) becomes rather smaller, $\approx 1.4$. Eq.\ (17) is a
729: conservative estimate of fluence uncertainty, though it still
730: omits other possible systematic uncertainties. Note that in the
731: EGRET era, GRB 930131 had the largest number of photons,
732: $N_\gamma(100$ MeV -- 5 GeV)$\approx 12$, $\gamma$ rays, in the
733: 0.1 -- 5 GeV range. As is apparent from Fig.\ \ref{fig3new},
734: finite photon number gives the major source of error to the
735: fluence measurement of EGRET spark-chamber GRB.
736:
737: In comparison, the error associated with relative energy
738: uncertainty is, as already noted, at the $\simeq 10$\% level for
739: EGRET. For the Fermi Gamma ray Space Telescope, the relative
740: energy uncertainties at photon energies of 0.1, 1, and 10 GeV are
741: $\approx 17$\%, 9\%, and 8\% for normal incidence, and $\approx
742: 15$\%, 9\%, and 5\% at 60$^\circ$ off-axis \citep{atw09}. Because
743: this can be a calibration uncertainty affecting all photon energy
744: measurements systematically rather than statistically, this
745: uncertainty on the fluence will start to dominate for those rare
746: GRBs detected with Fermi with more than $\sim 50$ photons between
747: 100 MeV and 5 GeV. Effective area uncertainties also appear at the
748: $\sim 10$\% level. Consequently, the uncertainty from the finite
749: number of photons dominates fluence measurement for EGRET GRBs.
750:
751: From the discussion of EGRET GRBs in Section 3.1 and their
752: properties in Table~\ref{tbl-2}, we can now reliably estimate the
753: fluence uncertainty associated with finite number of photons,
754: background and deadtime effects measured with EGRET in the 0.1 --
755: 5 GeV energy range. For GRB 910503, we take $N_\gamma(100$ MeV --
756: 5 GeV) = 2, essentially giving 100\% uncertainty in the
757: measurement of fluence (taking twice this value would give a
758: 2$\sigma$ upper limit). For GRB 910601, we again conservatively
759: take $N_\gamma(100$ MeV -- 5 GeV) = 2 in view of possible
760: background. For GRB 930131, we take $N_\gamma(100$ MeV -- 5 GeV) =
761: 12, keeping in mind that a deadtime correction in the absolute
762: fluence by tens of percent or more could be required, depending on
763: the distribution of $\gamma$-ray arrival times, which Fermi will
764: measure. For GRB 940217, we take $N_\gamma(100$ MeV -- 5 GeV) =
765: 8, subtracting two background photons that could have contributed
766: during the first prompt and early afterglow in this GRB (note that
767: no EGRET photons in the 100 MeV -- 5 GeV range were detected after
768: Earth occultation). For GRB 940301, we take $N_\gamma(100$ MeV --
769: 5 GeV) = 5.
770:
771: \subsection{Fluence Estimate from EGRET Photon Event List}
772:
773: The statistical errors associated with the finite number of
774: detected photons, being the largest source of error for the EGRET
775: GRBs, are used with the values in Table~\ref{tbl-2} and eq.\
776: (\ref{fluence}) to produce Fig.\ \ref{fig4new}. This figure
777: displays the 100 MeV -- 5 GeV fluence measured with the EGRET
778: spark chamber vs. the $20$ keV -- 2 MeV fluence measured with
779: BATSE. Bolometric corrections to a larger energy range are usually
780: small if $E_{\rm pk}$, the peak photon energy of the $\nu F_\nu$
781: spectrum, falls in this waveband. This is not always the case for
782: these GRBs, and comparison of this figure with reported results in
783: the Fermi era must consider bolometric corrections.
784:
785: \begin{figure}[t]
786: \includegraphics[width=3.50in]{f4.eps}
787: \caption{Diagram plotting fluence measured between 100 MeV and 5 GeV with
788: the EGRET spark chamber vs the 20 keV -- 2 MeV fluence measured with BATSE
789: for the 5 EGRET spark chamber GRBs. The EGRET fluence for GRB 940217 could easily
790: be a factor of two larger due to Earth occultation during the active phase of this
791: GRB.
792: \label{fig4new}}
793: \end{figure}
794:
795: This fluence-fluence diagram provides, potentially, one way to
796: discriminate between different classes of GRBs on the basis of
797: their $\gamma$-ray radiative efficiency. A much more accurate
798: comparison will be provided for those GRBs with known redshift.
799: Because deadtime and background effects have been corrected from
800: these EGRET spark chamber GRBs, the fluence-fluence diagram gives
801: a picture where the long-duration GRBs have a radiative efficiency
802: $\rho$ in $>100$ MeV $\gamma$ rays of $\approx 1$\% or a few \%
803: of the bolometric energy output. Assuming no class bias in the
804: values in Table~\ref{tbl-2}, these give the average deduced
805: radiative efficiency $\langle \rho \rangle \cong 9$\% (column 8).
806: This suggests that we use as a mean fluence ratio a number between
807: $\approx 10$, before considering broad ranges of value of $\rho$
808: or even separate classes with different $\rho$ values.
809:
810: \begin{deluxetable*}{lcccc}
811: \hskip-1.in
812: \tabletypesize{\scriptsize} %
813: \tablecaption{Fermi and AGILE GRBs/yr FOV ($\approx
814: 1/5^{th}$) Predictions %
815: \label{tbl-3}} %
816: \tablewidth{0pt} %
817: \tablehead{ %
818: \colhead{} & %
819: \multicolumn{2}{c}{spectral index of -2} & %
820: \multicolumn{2}{c}{spectral index of -2.2} \\ %
821: \colhead{Threshold Energy} & %
822: \colhead{$\rm Fermi $} & %
823: \colhead{$\rm AGILE $} & %
824: \colhead{$\rm Fermi $} & %
825: \colhead{$\rm AGILE $} %
826: } %
827: \startdata
828: $>$ 30 MeV \hskip+0.05in ($> 5$ photons) & %
829: 30$^a$ (18)$^b$ & %
830: 7$^a$ (4)$^b$ & %
831: 12$^c$ (6)$^d$ & %
832: 2$^c$ ($<$ 1)$^d$ %
833: \\
834: $>$ 100 MeV ($> 5$ photons) & %
835: \hskip-0.05in 20 \hskip+0.07in (12) & %
836: \hskip+0.1in 2 \hskip+0.05in ($<$ 1)& %
837: \hskip+0.05in 8 (4) & %
838: \hskip-0.12in $<$ 1 ($<$ 1) %
839: \\
840: $>$ 1 GeV \hskip+0.13in ($> 5$ photons) & %
841: \hskip+0.08in 4 \hskip+0.07in ($<$ 2) & %
842: \hskip-0.12in $<$ 1 (0) & %
843: \hskip-0.08in $<$ 1 (0) & %
844: \hskip-0.23in $<$ 1 (0) %
845: \enddata
846: \tablecomments{This column has value with fluence ratio of
847: $^a$10\%; $^b$5\%; $^c$8\%; $^d$4\%.}
848:
849: \end{deluxetable*}
850:
851: \begin{figure*}[t]
852: \includegraphics[width=7.in]{f5.eps}
853: \caption{Prediction for the number of long-duration GRBs per year
854: full-sky Fermi LAT (a and b) and AGILE (c and d) will see based on
855: a measured fluence ratio extrapolated from a $-2$ photon spectum
856: from EGRET/BATSE to EGRET LAT energies. The curves describe the
857: burst rate assuming a $-2$ photon spectra for different energy
858: thresholds.\\}
859: %
860: \label{f5}
861: \end{figure*}
862: %
863:
864: %
865: \begin{figure*}[t]
866: \includegraphics[width=7.in]{f6.eps}
867: \caption{Prediction for the total number of GRBs/yr full-sky that
868: Fermi LAT (a and b) and AGILE (c and d) will detect for GRB
869: spectra as described in Fig.~3 for an assuming $-2.2$ photon
870: spectra.\\}
871:
872: \label{f6}
873: \end{figure*}
874: %
875:
876: \section{Alternate Fluence Estimate}
877:
878: As a check on $\rho$, we also make an independent estimate using
879: the number $N_\gamma(>100$ MeV) of $>100$ MeV $\gamma$-ray photons
880: that were detected by EGRET \citep{din95}, assuming a hard
881: component of the spectrum with spectral indices $\alpha_1 = -2$
882: and $\alpha_1 = -2.2$. The EGRET fluence in the energy range of
883: 100 MeV to 5 GeV can be expressed as
884: %
885: \begin{equation}
886: \Phi_{\rm EGRET} = N_\gamma (> 100 {\rm~MeV})\;{
887: \int^{5 \,{\rm GeV}}_{100\, {\rm MeV}} E^{1-\alpha_1} dE \over
888: \int^{5\,{\rm GeV}}_{100\, {\rm MeV}} E^{-\alpha_1} A(E,\theta) dE
889: }\;,
890: \end{equation}
891: %
892: where $A(E,\theta)$ is the angle-dependent EGRET effective area
893: obtained by interpolating the data in Figure~14 of \citet{tho93}.
894: From Table~\ref{tbl-2} we find that the average measured fluence
895: ratios $\rho$ for the 5 spark chamber bursts are $\rho \approx
896: 9\%$ (column 10) and $\rho \approx 7\%$ (column 12) for spectra
897: indices of $\alpha_1 = -2$ and $-2.2$, respectively. This is
898: consistent with our previous estimate of $\approx 9$\%, because
899: that referred to the $> 100$ MeV--5 GeV fluence to 20 keV--2 MeV
900: fluence ratio, rather than the $> 30$ MeV--5 GeV fluence to 20
901: keV--2 MeV fluence ratio, the latter of which is expected to be
902: $\approx 50$\% larger. However, these average fluence values of
903: 9\% and 7\% are dominated by the boundary of long-soft and
904: short-hard GRB 930191 with the fluence ratios of 40\% and 30\% for
905: spectra indices of -2 and -2.2, respectively. Four other bursts
906: are long GRBs with the average fluence ratios of 1.3\% and 1\%.
907: Hence, we expect the average fluence for the long GRBs to be
908: larger, conservatively, we assume values of 5\% and 4\% for
909: spectra indices of $\alpha_1 = -2$ and $-2.2$, respectively, to
910: represent long-duration GRBs.
911:
912: However, as previously noted, the total $\gamma$-ray flux is
913: probably larger than EGRET measured. The $\gamma$-ray emission
914: from GRB 940217 \citep{hur94} likely persisted during the period
915: of Earth occultation, easily increasing the number of photons by
916: $\sim 2$. Deadtime effects, as seen in GRB 930131 \citep{som94},
917: could have caused the number of $\gamma$ rays to be undercounted.
918: Besides, there is the example of GRBs with anomalous hard tails,
919: like GRB 990104 \citep{wbr02} and GRB 941017 \citep{gon03}, the
920: latter with fluence in the $\approx 1$ -- $100$ MeV range
921: comparable to or greater than the sub-MeV fluence. Further, there
922: is the low-significance Milagrito detection of the X-ray
923: flash-type GRB 970417a \citep{atk03}, which would not only have to
924: be at low redshift $z \lesssim 0.3$ but also have much greater TeV
925: than MeV fluence. Hence we expect the average fluence ratios
926: $\rho$ between EGRET and BATSE to be $\approx 10$\% and $\approx
927: 8$\% for spectra indices of $\alpha_1 = -2$ and $\alpha_1 = 2.2$,
928: respectively.
929:
930: We use $\rho = 0.1$ for model with $\alpha_1 = -2$ and $\rho =
931: 0.08$ for model with $\alpha_ 1 = -2.2$ to define the value of
932: $k_G$. The estimated total number of long-duration GRBs that get
933: detected by any detectors are based on the BATSE 4B fluence
934: distribution \citep{pac99}, shown in Fig.~\ref{fig2}, assuming an
935: annual GRB rate of 670 GRB/yr, including untriggered GRBs with
936: photon fluxes smaller than 0.3 cm$^{-2}$ s$^{-1}$ \citep{ban02}.
937: The BATSE 4B Catalog contains 1292 bursts total, including short
938: and long duration GRBs. There are a total of 872 long-duration
939: GRBs that are identifiable in the BATSE 4B Catalog. When we
940: estimate the number of source counts over one year period, the
941: ratio 872/1292 is applied to represent the long-duration GRB
942: predictions.
943:
944: \section{Results}
945:
946: In Figs.\ \ref{f5} and \ref{f6}, we plot full-sky Fermi LAT and
947: AGILE GRB detection rates, assuming that the Band-spectrum extends
948: to $\gtrsim 100 $ MeV energies. This simple model predicts that
949: $\approx 30, 20$, and $4$ GRBs will be observed per year with the
950: Fermi LAT with at least 5 photons with energies $E > 30$ MeV, $E >
951: 100$ MeV, and $E> 1$ GeV, respectively. This is assuming a hard
952: spectral index $\alpha_1 = -2$ and a fluence ratio $\rho =
953: 10$\%. For $\alpha_1 = -2.2$, the model predicts that $\approx
954: 12$, $\approx 8$, and probably no GRBs per year should be detected
955: with at least 5 photons with $E > 30$ MeV, $E > 100$ MeV, and $E>
956: 1$ GeV, respectively, now assuming that $\rho = 8$\%.
957:
958: The analysis also predicts that $\approx 7$ and $\approx 2$ GRBs
959: per year with at least 5 photon counts at $E >$ 30 MeV and $E >$
960: 100 MeV, respectively, should be detected by AGILE assuming
961: $\alpha_1 = -2$ and $\rho =10$\% (see Fig.~5d). For $\alpha_1 =
962: -2.2$, the model predicts that $\approx 2$ and probably no GRBs
963: with at least 5 photons with $E > 30$ MeV and $E > 100$ MeV,
964: respectively, will be detected by AGILE per year with $\rho = 8$\%
965: (see Fig. 6d). A consistently smaller number of GRBs/yr full-sky
966: is predicted for both Fermi and AGILE for the steeper spectral
967: indices, as expected. In Table~\ref{tbl-3}, we summarize the
968: predicted number of GRBs per yr with at least 5 photons greater
969: than a specific $\gamma$-ray photon energy in the FoV, $\approx
970: 2.5$ sr, of the Fermi LAT and AGILE, considering GRBs with hard
971: spectral indices $\alpha_1 = -2$ and $\alpha_1 = -2.2$, and for
972: fluence ratios $\rho = 10$\%, 8\%, 5\%, and 4\%.
973:
974: These models assume that all GRBs have power law with $\alpha_1
975: \approx -2$ and 10\% fluence ratio or $-2.2$ and 8\% fluence
976: ratios, and these assumptions would be first to be examined if
977: Fermi LAT shows strong violations of these predictions. For
978: $\alpha_1 = -2$ and $\rho = 10$\%, we estimate from Fig.\ 5 that
979: $\approx 1$ -- 2 GRBs per year will be detected with the Fermi
980: telescope that have more than 100 $\gamma$ rays with $E\gtrsim
981: 100$ MeV. The models also predicts that the GBM fluences of
982: long-duration GRBs detected with the LAT should be large, $\gtrsim
983: 10^{-5}$ ergs cm$^{-2}$. The distribution of fluences can be
984: compared with predicted distributions for Swift and the Fermi GBM
985: \citep{ld07}.
986:
987: From this treatment of the fluence ratio, we can estimate the
988: diffuse extragalactic $\gamma$-ray background produced by GRBs
989: using the 4B BATSE fluence catalog based on the EGRET/BATSE
990: fluence ratio. From the 4B Catalog there are a total of 1292 GRBs,
991: of which 872 bursts are long-duration GRBs. The total average
992: fluence for the long-duration GRB is about $\approx 1.40 \times
993: 10^{-2} \rm ~ergs~cm^{-2}$. Assuming 670 GRBs/yr, we estimate that
994: less than 1\% of the diffuse extragalactic $\gamma$-ray background
995: could come from GRBs assuming $\approx 8\%$ or 10\% fluence ratio.
996: Unless there is a large class of $\rho\approx 1$ GRBs, which were
997: not found with the EGRET spark chamber, this suggests that GRBs
998: gives very little contribution to the diffuse extragalactic
999: $\gamma$-ray background \citep[e.g.,][]{der06,cdz07}.
1000:
1001: \section{Conclusions}
1002:
1003: As we enter the Fermi era, it is worthwhile to make final
1004: predictions based on results from the EGRET CGRO era. Here we have
1005: adopted an approach making use of mean fluence ratios in the EGRET
1006: 100 MeV -- 5 GeV and BATSE 20 keV -- 2 MeV bandpasses. Obviously
1007: this already introduces significant K-corrections when calculating
1008: the intrinsic properties of the GRB sources; a sample of GRBs with
1009: known redshifts is clearly preferred.
1010:
1011: Nevertheless, we fix the fluence ratios in these bands to obtain
1012: straightforward predictions for all-sky detection rates with
1013: Fermi, using on-axis effective areas and the BATSE 4B fluence
1014: distribution. Multiplication by $1/5^{\rm th}$ gives the final
1015: detection rate with the Fermi LAT tracker and calorimeter
1016: \citep{atw09}
1017:
1018: For a 10\% fluence ratio and a hard component with $-2$ index, we
1019: estimate $\approx 1$ -- 2 per month with more than 5 photons with
1020: $E> 100$ MeV $\gamma$ rays in the LAT FoV, and 1 or 2 per year
1021: with more than 100 $\gamma$ rays with $E > 100$ MeV. A bright GRB
1022: for spectral modeling, with more than $\approx 100$ photons with
1023: $E>100$ MeV, is predicted once or twice per year. We note that
1024: with the use of $-2$ fluence spectrum, the detection of a $> 10$
1025: GeV photon from a GRB is considered very improbable, barring the
1026: existence of an additional component harder than $-2$.
1027:
1028: Furthermore, as the Fermi GBM and the Fermi LAT instruments begin
1029: to collect data from GRBs, we recommend to plot LAT fluence vs.\
1030: GBM fluence and, more importantly, mean LAT GRB spectral index
1031: vs.\ LAT/GBM fluence ratios $\rho$ to search for distinct classes
1032: of GRBs distinguished by their high-energy properties.
1033:
1034: \acknowledgements We thank David Band, Michael Briggs, Valerie
1035: Connaughton, Brenda Dingus, Magda Gonz\'alez, Chryssa Kouveliotou,
1036: Julie McEnery and Nicola Omodei for many interesting discussions
1037: and useful suggestions, and especially to Nicola Omodei and Julie
1038: MeEnery for a really difficult and constructive report. We also
1039: thank the referee for comments, corrections, and useful
1040: suggestions. The work of CD is supported by the Office of Naval
1041: Research and NASA Fermi Science Investigation DPR-S-1563-Y, which
1042: also supported the research of TL at the Naval Research
1043: Laboratory.
1044:
1045: \begin{thebibliography}{}
1046:
1047: \bibitem[Atkins et al.(2003)]{atk03} Atkins, W.\ B., et al.\
1048: 2003, \apj, 583, 824
1049:
1050: \bibitem[Atwood et al.(2009)]{atw09} Atwood, R., et al.\
1051: 2009, \apj, to be submitted
1052:
1053: \bibitem[Band et al.(1993)]{ban93} Band, D., et al.\ 1993,
1054: \apj, 413, 281
1055:
1056: \bibitem[Band(2002)]{ban02} Band, D.~L.\ 2002, \apj, 578, 806
1057:
1058: \bibitem[Baring(2006)]{bar06} Baring, M.~G.\ 2006, \apj, 650,
1059: 1004
1060:
1061: \bibitem[Belli et al.(2007)]{bel07} Belli, F. et al.\ 2007,
1062: Nuclear Instruments and Methods in Physics Research, A570, 276
1063:
1064: \bibitem[Briggs et al.(1999)]{bri99} Briggs, M.~S., et al.\
1065: 1999, \apj, 524, 82
1066:
1067: \bibitem[Casanova et al.(2007)]{cdz07} Casanova, S., Dingus,
1068: B.~L., \& Zhang, B.\ 2007, \apj, 656, 306
1069:
1070: \bibitem[Catelli et al.(1998)]{cds98} Catelli, J.~R., Dingus,
1071: B.~L., \& Schneid, E.~J.\ 1998, American Institute of Physics Conference
1072: Series, 428, 309
1073:
1074: \bibitem[Dermer(2006)]{der06} Dermer, C. D., 2006, in The Multi-Messenger Approach to High-Energy
1075: Gamma-Ray Sources, eds. J.\ M.\ Paredes, O.\ Reimer, and D.\ F.\
1076: Torres (New York: Springer), 127
1077:
1078: \bibitem[Dermer(2007)]{der07} Dermer, C. D. 2007, \apj, 659, 958
1079:
1080: \bibitem[Dermer \& Dingus(2004)]{dd04} Dermer, C.~D., \&
1081: Dingus, B.~L.\ 2004, New Astr.\ Rev., 48, 537
1082:
1083: \bibitem[Dingus(1995)]{din95} Dingus, B.~L.\ 1995, \apss,
1084: 231, 187
1085:
1086: \bibitem[Dingus et al.(1998)]{din98}
1087: Dingus, B.~L., Catelli, J.~R., \& Schneid, E.~J.\ 1998, American
1088: Institute of Physics Conference Series, 428, 349
1089:
1090: \bibitem[Gonz{\'a}lez et al.(2003)]{gon03} Gonz{\'a}lez,
1091: M.~M., Dingus, B.~L., Kaneko, Y., Preece, R.~D., Dermer, C.~D., \& Briggs,
1092: M.~S.\ 2003, \nat, 424, 749
1093:
1094: \bibitem[Gonz{\'a}lez et al.(2009)]{gon09} Gonz{\'a}lez,
1095: M.~M., et al.\ 2009, \apj, submitted
1096:
1097: \bibitem[Hurley et al.(1994)]{hur94} Hurley, K., et al.\
1098: 1994, \nat, 372, 652
1099:
1100: \bibitem[Kaneko et al.(2008)]{kgp08} Kaneko, Y.,
1101: Gonz{\'a}lez, M.~M., Preece, R.~D., Dingus, B.~L., \& Briggs,
1102: M.~S.\ 2008, \apj, 677, 1168
1103:
1104: \bibitem[Kippen et al.(2001)]{kip01} Kippen, R.~M., et al.\
1105: 2001, Gamma 2001: Gamma-Ray Astrophysics, 587, 801
1106:
1107: \bibitem[Kouveliotou et al.(1993)]{kou93} Kouveliotou, C.,
1108: Meegan, C.~A., Fishman, G.~J., Bhat, N.~P., Briggs, M.~S., Koshut, T.~M.,
1109: Paciesas, W.~S., \& Pendleton, G.~N.\ 1993, \apjl, 413, L101
1110:
1111: \bibitem[Kouveliotou et al.(1994)]{kou94} Kouveliotou, C., et
1112: al.\ 1994, \apjl, 422, L59
1113:
1114: \bibitem[Kulkarni et al.(1999)]{kul99} Kulkarni, S.~R., et
1115: al.\ 1999, \nat, 398, 389
1116:
1117: \bibitem[Kwok et al.(1993)]{kwo93} Kwok, P.~W., et al.\ 1993,
1118: American Institute of Physics Conference Series, 280, 855
1119:
1120: \bibitem[Le \& Dermer(2007)]{ld07} Le, T., \& Dermer, C.~D.\
1121: 2007, \apj, 661, 394
1122:
1123: \bibitem[Meegan et al.(1996)]{mee96} Meegan, C.~A., et al.\
1124: 1996, \apjs, 106, 65
1125:
1126: \bibitem[Merck et al.(1995)]{mer95} Merck, M., et al.\ 1995,
1127: IAU Colloq.~151: Flares and Flashes, 454, 358
1128:
1129: \bibitem[Norris
1130: \& Gehrels(2008)]{ng08} Norris, J.~P., \& Gehrels, N.\ 2008, American
1131: Institute of Physics Conference Series, 1000, 280
1132:
1133: \bibitem[Omodei (2006)]{omo06}
1134: Omodei, N., \& for the GLAST/LAT GRB Science Group 2006,
1135: astro-ph/0603762
1136:
1137: \bibitem[Paciesas et al.(1999)]{pac99} Paciesas, W.~S., et
1138: al.\ 1999, \apjs, 122, 465
1139:
1140: \bibitem[Pittori et al.(2004)]{pit04} Pittori, C., Tavani,
1141: M., \& AGILE Team 2004, Bulletin of the American Astronomical
1142: Society, 36, 929
1143:
1144: \bibitem[Preece et al.(2000)]{pre00} Preece, R.~D., Briggs,
1145: M.~S., Mallozzi, R.~S., Pendleton, G.~N., Paciesas, W.~S., \&
1146: Band, D.~L.\ 2000, \apjs, 126, 19
1147:
1148: \bibitem[Schneid et
1149: al.(1992)]{sch92} Schneid, E.~J., et al.\ 1992, \aap, 255, L13
1150:
1151: \bibitem[Schneid et al.(1995)]{sch95} Schneid, E.~J., et al.\
1152: 1995, \apj, 453, 95
1153:
1154: \bibitem[Sommer et al.(1994)]{som94} Sommer, M., et al.\
1155: 1994, \apjl, 422, L63
1156:
1157: \bibitem[Sreekumar et al.(1998)]{sre98} Sreekumar, P., et
1158: al.\ 1998, \apj, 494, 523
1159:
1160: \bibitem[Stecker et al.(2008)]{shk08} Stecker, F.~W., Hunter,
1161: S.~D., \& Kniffen, D.~A.\ 2008, Astroparticle Physics, 29, 25
1162:
1163: \bibitem[Strong et al.(2004)]{str04} Strong, A.~W.,
1164: Moskalenko, I.~V., \& Reimer, O.\ 2004, \apj, 613, 956
1165:
1166: \bibitem[Thompson et al.(1993)]{tho93} Thompson, D.~J., et
1167: al.\ 1993, \apjs, 86, 629
1168:
1169: \bibitem[Thompson(1986)]{tho86} Thompson, D.~J.\ 1986,
1170: NIM Physics Research A, 251, 390
1171:
1172: \bibitem[Wren et al.(2002)]{wbr02} Wren, D.~N., Bertsch,
1173: D.~L., \& Ritz, S.\ 2002, \apjl, 574, L47
1174:
1175: \end{thebibliography}{}
1176:
1177: \end{document}
1178: