1: \documentclass[preprint]{emulateapj}
2: %\documentclass[12pt,preprint]{aastex}
3:
4: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
5: % Blazar paper II
6: % Pavlidou & Venters
7: %
8: % first version
9: %
10: % written in aastex
11: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
12:
13: \newcommand\beq{\begin{equation}}
14: \newcommand\eeq{\end{equation}}
15: \newcommand\beqar{\begin{eqnarray}}
16: \newcommand\eeqar{\end{eqnarray}}
17:
18: \newcommand\ee[2]{#1 \times 10^{#2}}
19:
20: \newcommand\pref[1]{(\ref{#1})}
21: \newcommand\pcite[1]{(\cite{#1})}
22: \newcommand\etal{{et al.~}}
23:
24: \newcommand\pion[1]{\mbox{$\pi^{#1}$}}
25:
26: \newcommand\msol{\hbox{$M_{\odot}$}}
27: \newcommand\lesc{\ell_{\rm esc}}
28: \newcommand\OmegaM{\Omega_{\rm m}}
29: \newcommand\ngdot{\dot{n}_{\gamma,{\rm com}}}
30:
31: \newcommand{\flf}{\mathcal{F}}
32:
33: %horizontal line for repeated author entry in bibliography
34: \newcommand\bline{\rule[1.2mm]{3em}{0.1mm}}
35:
36: %SN rate symbol
37: \newcommand\snr[1]{\mbox{${\cal R}_{_{\rm #1}}$}}
38:
39: % Intensity (flux per solid angle) units
40: \newcommand\iun{\mbox{${\rm \, photons \,\, cm^{-2} \, s^{-1}\,sr^{-1}}$}}
41:
42: % Flux units
43: \newcommand\fun{\mbox{${\rm \, photons \,\, cm^{-2} \, s^{-1} \, sr^{-1}}$}}
44:
45: \begin{document}
46:
47: \title{The Spectral Shape of the Gamma-ray Background from Blazars }
48:
49: \author{Vasiliki Pavlidou\altaffilmark{1} \&
50: Tonia M. Venters\altaffilmark{2,3} }
51:
52:
53: \altaffiltext{1}{
54: Kavli Institute of Cosmological Physics and Enrico Fermi Institute,
55: The University of Chicago, Chicago, IL 60637}
56: \altaffiltext{2}{Department of Astronomy and Astrophysics, The University
57: of Chicago, Chicago, IL 60637}
58: \altaffiltext{3}{Laboratoire d'AstroParticule et Cosmologie, Universit{\'e} de Paris 7 - Denis Diderot, Paris 13$^{\mbox{\`e}me}$, France}
59:
60: \begin{abstract}
61:
62: The spectral shape of the unresolved emission from different classes
63: of gamma-ray emitters can be used to disentangle the contributions
64: from these populations to the extragalactic gamma-ray background
65: (EGRB).
66: We present a calculation of the unabsorbed {\em spectral shape} of the
67: unresolved blazar contribution to the EGRB starting from
68: the spectral
69: index distribution (SID)
70: of resolved EGRET blazars derived through a maximum-likelihood
71: analysis accounting for measurement errors. In addition, we explicitly
72: calculate the {\em uncertainty} in this theoretically predicted
73: spectral shape, which enters through the spectral index distribution
74: parameters. We find that: (a) the
75: unresolved blazar emission spectrum is only mildly convex, and thus, even
76: if blazars are shown by GLAST to be a dominant contribution to the
77: ERGB at lower energies, they may be insufficient to explain the EGRB
78: at higher energies; (b) the theoretically predicted unresolved
79: spectral shape involves significant uncertainties due to the limited
80: constraints provided by EGRET data on the SID parameters, which are comparable to the
81: statistical uncertainties of the observed EGRET EGRB at high
82: energies; (c) the increased number statistics which will be provided by
83: GLAST will be sufficient to reduce this uncertainty by at least a
84: factor of three.
85: \end{abstract}
86: \keywords{galaxies: active -- gamma rays: observations -- gamma rays:
87: theory -- diffuse radiation}
88:
89: \maketitle
90:
91: \section{Introduction}
92:
93: The isotropic, and presumably, extragalactic gamma-ray background
94: emission (EGRB) detected by the Energetic Gamma-ray Experiment
95: Telescope (EGRET) aboard the {\it Compton Gamma-ray Observatory}
96: (Sreekumar et al.\ 1998) is one of the most important observational
97: constraints on known or theorized populations of faint, unresolved
98: gamma-ray emitters. With the imminent launch of the {\it Gamma-ray
99: Large Area Space Telescope} (GLAST), which is expected to represent
100: an unprecedented leap in observational capabilities in GeV energies,
101: the timing is especially opportune to consider the information content
102: of the diffuse background and methods for maximizing the scientific
103: return from its study.
104:
105: One of the primary challenges in using the EGRB to constrain
106: properties of extragalactic gamma-ray emitters and exotic physics is
107: disentangling the convolved contributions of guaranteed participating
108: populations. Estimates of the levels of the collective unresolved
109: emission even from established classes of extragalactic sources (such
110: as blazars and normal galaxies) involve significant uncertainties and
111: are at the order-of-magnitude level at best (e.g., Padovani et al.\
112: 1993; Stecker \& Salamon 1996; Kazanas \& Perlman 1997; Mukherjee \&
113: Chiang 1999; M\"ucke \& Pohl 2000; Dermer 2006; Lichti et al.\ 1978;
114: Pavlidou \& Fields 2002).
115:
116: A very promising approach for the study of the EGRB and its components
117: is through the use of spectral shape information. Let us consider the
118: optimal case where the expected spectral shapes of the unresolved
119: emission from known classes of gamma-ray sources can be confidently
120: predicted. In this case, a series of conclusions can be drawn
121: regarding the potential contributions of these classes to the EGRB
122: even without detailed calculations of the {\em magnitudes} of their
123: collective emission.
124: For example, in comparing the spectral shape of the spectrum due to a
125: particular class with that of the EGRB, one can identify whether this
126: class could, in principle, comprise most of the EGRB or require the
127: existence of contributions from other classes (Stecker \& Salamon
128: 1996a,b; Strong et al. 2004; Pavlidou \& Fields 2002)\footnote{Note
129: that the method of spectral comparison can only be used to {\em
130: reject} a population from being the sole source of the EGRB;
131: spectral consistency does not constitute in itself proof of the
132: importance of a class of objects as an EGRB contributor, since the
133: overall normalization of the emission may, in fact, be low depending on the gamma-ray luminosity function of the population.}.
134: Potentially
135: identifiable spectral features could be predicted and searched for in
136: GLAST data (e.g., Pavlidou \& Fields 2002). Finally, spectral
137: information could be used to ultimately disentangle different
138: components and contributions (as in e.g. de Boer et al.\ 2004 for the
139: case of the diffuse emission from the Milky Way). An additional
140: attractive feature of such calculations is that the associated
141: uncertainties are largely independent of those entering the
142: calculations of the overall unresolved emission flux.
143:
144: As blazars are the most populated class of gamma-ray emitters, unresolved blazars are guaranteed to contribute significantly, if not dominantly, to the EGRB. Thus, it is especially important to understand the expected spectral shape of their collective unresolved emission and the uncertainties involved in its calculation.
145: Individual blazars have been measured to have power-law spectra in the EGRET range, $F_E \propto E^{-\alpha}$, with spectral index $\alpha$ ranging approximately from 1.5 to 3. The unresolved emission from a collection of power-law emitters with variable spectral indices\footnote{Statistically significant spread in the observed and intrinsic spectral index of blazars has also been confirmed in other energy bands (see e.g.\ Shen et al.\ 2006 for the case of X-ray emission).} has, invariably, a convex spectral shape (Brecher \& Burbidge 1972; Stecker \& Salamon 1996; Pohl et al.\ 1997; Pavlidou et al.\ 2007).
146: The exact shape of the unresolved spectrum depends on the spectral index distribution (SID) of gamma-ray loud blazars.
147: Recognizing that this is the case, Stecker \& Salamon (1996a) explicitly reconstructed a spectrum from the observed SID of blazars in the 2nd EGRET catalog (Thompson et al.\ 1995), deriving a spectrum that was indeed significantly convex. However, measurement errors in individual spectral indices smear the SID and exaggerate the curvature of the spectrum (Pohl et al.\ 1997).
148:
149: Recently, in Venters \& Pavlidou 2007 (hereafter VP07), we have
150: applied a maximum-likelihood analysis to recover the intrinsic
151: spectral index distribution (ISID) of gamma-ray loud blazars from
152: EGRET observations. We found that (1) the maximum-likelihood ISID is
153: appreciably narrower than the observed SID, so the
154: best-guess spectrum is likely to have only a mild curvature;
155: (2) BL Lacs and flat
156: spectrum radio quasars (FSRQs) are likely to be spectrally distinct
157: populations with spectrally distinct contributions to the EGRB; (3)
158: there is no evidence for a systematic shift of spectral variability
159: with flaring implying that although variability may be important in
160: the level of the contribution from blazars, ignoring variability
161: effects in spectral shape studies is likely to be a good
162: approximation.
163:
164: Here, we use the ISIDs derived in VP07 to calculate the spectral shape
165: of the blazar contribution to the EGRB. We examine the sensitivity of
166: the shape to the exact values of the ISID parameters and report on the
167: range of possible shapes given our uncertainties in the determination
168: of these parameters. We also investigate how the spectral shapes of
169: the BL Lac and FSRQ contributions may differ. Finally, we predict how
170: our understanding of the spectral shape of the unresolved blazar
171: emission will improve after GLAST observations become available.
172:
173: \section{Formalism}
174:
175: If the differential photon flux spectrum of a single blazar is
176: $F_E(E) = F_{E,0} (E/E_0)^{-\alpha}$ (photons per unit area per unit energy per unit time), then the total flux of photons with
177: energies $>E_0$ is $F(>E_0) = F_{E,0}E_0 /(\alpha-1)$ (photons per
178: unit area per unit time). The
179: contribution of a single unresolved blazar of $F(>E_0) =F$ to the
180: EGRB is
181: \begin{equation}
182: I_{1} = (\alpha-1)\frac{F}{4\pi E_0}\left(\frac{E}{E_0}\right)^{-\alpha}
183: \end{equation}
184: where $I$ has units of photons per unit area per unit energy per unit time per unit solid
185: angle, and the flux of one source is uniformly distributed over
186: $4\pi$.
187:
188: Now let us assume that the flux distribution of unresolved blazars
189: (number of objects per flux interval) can be described by a function
190: $g(F)$; and that the distribution of spectral indices (number of
191: objects per spectral index interval) can be described by a function
192: $p(\alpha)$. Then the total contribution of unresolved blazars to the
193: EGRB is
194: \begin{equation}\label{complex}
195: I_{\rm EGRB}(E) =
196: \int_{\alpha = -\infty}^{\infty}d\alpha
197: \int_{F=0}^{F_{\rm min}}\!\!\!\!\!dF \, g(F) I_{1} p(\alpha)\,. \nonumber \\
198: \end{equation}
199: In Eq.\ (\ref{complex}), $F_{\rm min}$ is the minimum flux of an object
200: that can be resolved by the telescope under consideration.
201:
202: We now make the assumptions that (a) blazar spectra can be adequately described by single power-laws in the observed energy range, as well as at energies which redshift down to the observed range; (b) the flux distribution is independent of spectral index; and (c) the spectral index does not evolve with time and does not depend on luminosity.
203: %\footnote{Note that this assumption is meaningful only for
204: %the ISID and NOT for the luminosity function.}.
205: In this case, blazars in any single flux interval sample an identical SID, and produce the same fraction of photons in any two energy bins, thus resulting in a unique spectral shape. This is intuitively reasonable. Had the (non-evolving) ISID been a $\delta$-function, all blazars would be power laws of the same slope, independently of the epoch of observation and the coadded spectrum would identically be a power law of the same slope. Since our ISID is a Gaussian of non-evolving spectral indices, the spectral shape is curved, but still independent of the luminosity function. The blazars in a specific unresolved flux interval contributing to the background will represent a mixture of luminosities and redshifts, but since the blazar properties that determine the flux interval (redshift, luminosity) do not depend on the spectral index, the blazars within that flux interval will fairly sample the same ISID. Additionally, redshifting down the spectra has no effect on the slope for single power laws as long as no absorption occurs. The same reasoning is applicable to all flux intervals, which therefore contribute portions of the background that have different amplitude but the same, unique spectral shape, dependent only on the parameters of the ISID.
206:
207: The magnitude and spectral shape factors in Eq.\ (\ref{complex}) therefore decouple under our assumptions, and Eq.\ (\ref{complex}) can be rewritten
208: as
209: \begin{equation}\label{tria}
210: I_{\rm EGRB}(E) = I_0 \int_{\alpha = -\infty}^{\infty}\!\!\!\!d\alpha
211: (\alpha-1)\left(
212: \frac{E}{E_0}\right)^{-\alpha}\!\!\!\!\!p(\alpha)\,,
213: \end{equation}
214: where $I_0$ is a normalization constant depending on the flux
215: distribution of unresolved blazars.
216: If $p(\alpha)$ is Gaussian, as assumed in VP07, then Eq.\ (\ref
217: {tria}) is analytically integrable (Pavlidou et al.\ 2007).
218:
219: It should be noted that if the above assumptions do not hold, then the spectral shape may not decouple from its magnitude as above. This is particularly true if features are present in blazar spectra (breakdown of assumption a). However, there is no evidence in EGRET data for such features. In the cases where the other two assumptions (b and c) do not hold, as long as the dependencies are small, the decoupling of the shape and magnitude will still be approximately correct. In VP07, possible correlations between spectral index and redshift and between spectral index and luminosity for blazars were investigated and no evidence for such correlations was found. Nevertheless, absence of evidence is not equivalent to evidence of absence of a correlation, and it cannot, as yet, be proven that the spectral index is independent of redshift and luminosity. Since the gamma-ray emission from blazars is likely due to Inverse Compton emission which is related to emission at lower frequencies, a relationship between the gamma-ray spectral index and the gamma-ray luminosity is plausible and can be motivated, for example, in the context of the blazar sequence (Fossati et al. 1998; Ghisellini et al. 1998). On the other hand, the blazar sequence has been subsequently called into question with data from deeper blazar surveys (Giommi et al. 2005; Padovani 2007). Thus, it is difficult to interpret the VP07 result, especially based on current data alone; a further investigation of the blazar sequence (as suggested in Maraschi et al. 2007) as well as posible correlations of luminosity with spectral index in the GeV energy range with GLAST data will offer further insight into the issue. However, we point out that systematic uncertainties in the spectral shape of the blazar contribution entering through correlations between blazar spectral index and blazar luminosity/redshift {\em weaker} than the VP07 constraints would be dominated by the uncertainties entering through our limited knwoledge of the ISID, or the systematic uncertainties in the observational determination of the EGRB.
220:
221: For now, we can instill more confidence in the reasoning behind the above arguments by applying our formalism to an independent SID for which a shape was determined using the full redshift-dependent, luminosity-dependent equation.
222: %\footnote{Of course, for this case, magnitude was also determined.}.
223: We applied Eq. \ref{tria} to the Stecker \& Salamon (1996a) SIDs\footnote{Note that appropriate adjustments had to be made: (1) separation between flaring and quiescent blazars as they did, and (2) rejection of unphysical spectral indices arising from a wide SID.} and found the shapes for flaring and quiescent blazars reported by the authors (although, of course, our formalism cannot reproduce the {\em amplitude} of the emission they reported).
224:
225: \section{Inputs}
226:
227: Traditionally, in order to evaluate the blazar contribution to the
228: EGRB, one would derive a gamma-ray luminosity function and a redshift
229: distribution and calculate the overall \emph{magnitude} of the
230: contribution. However, in this analysis, we seek information about
231: the \emph{shape} of the blazar contribution rather than the overall
232: magnitude. Thus, our only inputs are the ISIDs for different EGRET samples and simulated GLAST samples presented in VP07. Since we do not include information about the magnitude of the blazar contribution to the EGRB, we have normalized our curves so that they always pass through the $83 {\rm \,
233: MeV}$ best-guess measurement of the EGRB from EGRET data
234: ($\left.E^2I_{\rm EGRB}\right|_{83 {\rm \, MeV}} = 1.5 \times 10^{-6}
235: {\rm \, GeV \, cm^{-2} s^{-1} sr^{-1}}$), as analyzed by Strong et al
236: (2004), in order to aid the visual comparison with the observed
237: EGRB shape. This normalization is completely arbitrary as we could have instead chosen any one of the Strong et al. 2004 data points; we have chosen $83 \mbox{ MeV}$ since at this energy the observational
238: statistical and systematic uncertainties are relatively low.
239:
240: %In order to determine how much differing ISIDs affect the spectral
241: %%shapes of the blazar contributions to the EGRB, we employ the ISIDs
242: %%%found in VP07 for FSRQs and BL Lacs in this analysis.
243: %%%Additionally, %we sample the parameter spaces within the 1$\sigma$
244: %%%contours of each %population to estimate the error on the spectral
245: %%%shape.
246:
247: \section{Results}
248:
249: \begin{figure}
250: %\resizebox{3.0in}{!}
251: %{\includegraphics{mattoxWithSymbols.eps}}
252: \plotone{f1.eps}
253: \caption{\label{fig1}
254: Spectral shape of the unresolved blazar emission. All curves have been normalized (arbitrarily) so that they pass through the 83 MeV point of the Strong et al.\ (2004) EGRB. Solid line: best-guess
255: spectrum based on a maximum-likelihood ISID determined using the Mattox et al.\ 2001 confident blazar sample. Grey region: spectral
256: shapes allowed for ISID parameters within the $1\sigma$ likelihood
257: contour. Filled triangles: Strong et al.\ (2004) EGRB
258: determination. Open circles: Sreekumar et al.\ (1998) EGRB
259: determination. Error bars are statistical errors only. Thick dotted
260: lines: Strong et al.\ (2004) EGRB systematics. Thin dotted line: SID determined in Stecker \& Salamon (1996a). }
261: \end{figure}
262:
263: The spectral shape of the unresolved blazar emission for the ISID of
264: the Mattox et al.\ (2001) confident blazars sample (46 sources) is
265: plotted in Fig.\ \ref{fig1}. The solid line is the best-guess
266: spectrum based on the maximum-likelihood Gaussian ISID for this sample
267: (mean $\alpha_0 = 2.27$ and spread $\sigma_0 = 0.2$) from VP07. The
268: gray region represents spectral shapes derived from $(\alpha_0,
269: \sigma_0)$ pairs within the 1$\sigma$ contour of the ISID parameter
270: likelihood and illustrates our $1\sigma$ uncertainty in the unresolved
271: blazar spectral shape. For comparison, we also plot datapoints with
272: (statistical) errors for the Sreekumar et al.\ (1998) (open circles)
273: and the Strong et al.\ (2004) (filled triangles)\footnote{The
274: Sreekumar et al.\ (1998) and the Strong et al.\ (2004) EGRB
275: determinations differ in the model used to subtract the diffuse
276: emission of the Milky Way with the difference resulting from the fact that the Strong et al.\ (2004) determination
277: is based on a Milky Way model which accounts for the GeV excess.}
278: determinations of the EGRB from EGRET data. We stress again that our
279: results are intended to be compared to the EGRET data only in shape
280: (relative intensity between different energy bins) and not in
281: amplitude, since the normalization of our curves does not carry any
282: information.
283: The thick dotted lines
284: indicate the Strong et al.\ (2004) systematics, entering through their
285: model of the Galaxy.
286: The best-guess spectrum is only mildly curved in comparison to the
287: best-guess spectrum derived from the observed,
288: measurement-error--contaminated SID (as parametrized by Stecker \&
289: Salamon 1996a for the Mattox et al.\ 2001 confident blazar dataset;
290: shown as the thin dashed line in Fig.\ \ref{fig1}) in lieu of the
291: maximum-likelihood ISID.
292: %Notably, the difference between the spectrum resulting from the
293: %%observed SID and that derived from the ISID is most pronounced at
294: %%%higher energies.
295:
296: \begin{figure*}
297: \begin{center}
298: \resizebox{4.5in}{!}
299: %{\includegraphics{Fourplot.eps}}
300: {\plotone{f2.eps}}
301: \caption{\label{fig2}
302: Predicted spectral shapes for FSRQs (left column) and BL Lacs (right column). All curves have been normalized (arbitrarily) to the 83 MeV point of the Strong et al.\ EGRB. Solid lines represent the best-guess spectra and the gray regions, the $1\sigma$ spectral shape uncertainties. Data points: Strong et al. (2004) EGRB determination (error bars represent statistical errors). Upper panel: EGRET data. Lower
303: panel: reduction of shape uncertainties in the GLAST era based on
304: the Dermer (2007) predictions for the numbers of detectable BL Lacs
305: and FSRQs. }
306: \end{center}
307: \end{figure*}
308:
309: A most important result is that the
310: constraints on the theoretical cumulative blazar spectrum are not
311: very strong. In fact, the theoretical uncertainties in the spectrum
312: entering through our limited understanding of the ISID are comparable
313: with the statistical uncertainties (errors on individual data points)
314: in the observed EGRB. Worse yet, the large systematic and statistical
315: uncertainties in the determination of the EGRB impede any comparison
316: between the theoretical cumulative blazar spectrum and the EGRB.
317: For this reason, no strong conclusions regarding the spectral
318: (in)consistency of the blazar collective spectrum with the observed
319: EGRB should be drawn based solely on EGRET data. Note that improvements in observations will \emph{not} automatically alleviate this concern. Even if the systematic and statistical uncertainties in the data improve, if the uncertainties in the theoretical spectral shape remain as they are, strong conclusions about the blazar contribution to the EGRB will remain unattainable. However, if we were to
320: ignore systematics and take at face-value the upturn of the EGRET EGRB
321: at high energies indicated by Strong et al.\ (2004), our results
322: suggest that it is unlikely that EGRET blazars, as a population,
323: comprise the dominant contribution to the background at the highest
324: energies even if they do dominate at low energies.
325:
326: If BL Lacs and FSRQs have distinct intrinsic spectral index
327: distributions as indicated tentatively by EGRET data (Pohl et al.\
328: 1997; Mukherjee et al.\ 1997; VP07), their cumulative unresolved
329: spectra will also differ. The top row of Fig.\ \ref{fig2} shows the
330: spectral shapes of unresolved emission from FSRQs (left panel) and BL
331: Lacs (right panel) calculated from the respective VP07 ISIDs derived
332: from EGRET data. The solid lines again represent best-guess spectra
333: and the gray regions are $1\sigma$ uncertainties in the shape. The
334: observed EGRB shape (Strong et al.\ 2004 with statistical error bars)
335: is indicated with the crosses.
336:
337:
338: The best-guess spectra of the two populations have different shapes
339: with the cumulative emission from BL Lacs being generally harder than
340: that of FSRQs and having a more convex spectrum. However, due to the
341: poor number statistics of BL Lacs, their ISID is not well
342: constrained and this uncertainty is carried over to the emission
343: spectrum: the theoretical uncertainties in the unresolved BL Lac
344: spectral shape are much larger than the statistical uncertainties in
345: the data and comparable with the systematic uncertainties in the
346: observed EGRB. Thus, no confident statement can be made regarding the
347: spectral (in)consistency of BL Lacs with the observed EGRB. On the
348: other hand, as there are many more FSRQs than BL Lacs detected by
349: EGRET, when all blazars are treated as a single population, both the
350: ISID and the unresolved emission spectrum closely resemble that of
351: FSRQs.
352:
353: If a similar ratio of FSRQs to BL Lacs is also present in the
354: unresolved blazar population (e.g., Dermer 2007) as in the resolved
355: EGRET blazar population, the shape of the blazar EGRB component will
356: mostly resemble that of Fig.\ \ref{fig1}. If on the other hand the
357: BL Lac fraction is much higher in unresolved blazars (e.g. Pohl et
358: al.\ 1997), then there may be an appreciable BL Lac contribution to
359: the EGRB, accounting, at least in part, for the upturn of the observed
360: EGRB tentatively suggested by EGRET data. GLAST observations will
361: greatly help in addressing this question as it is expected to resolve
362: between 1000 and 10,000 blazars, thus placing much stronger
363: constraints on the luminosity function and evolution of FSRQs and BL
364: Lacs.
365:
366: GLAST observations will also allow a much more confident determination
367: of the FSRQ and BL Lac ISIDs resulting in corresponding improvements
368: in the determinations of the unresolved emission spectral shapes for these
369: populations. The lower row in Fig.\ \ref{fig2} shows the improvements
370: of our theoretical predictions for the FSRQ (left panel) and BL Lac
371: (right panel) unresolved spectra using ISIDs from the simulated GLAST
372: datasets of VP07. Note that this is simply a prediction of
373: how much the uncertainties in the
374: determinations of the spectral shapes will be reduced
375: with increased number statistics and is not an actual prediction of
376: the GLAST EGRB. Shape uncertainties in the GLAST era are reduced by a
377: factor of $\sim 3$. FSRQs and BL Lacs were assumed to follow the Dermer
378: (2007) luminosity functions which represent the most conservative
379: (lowest) predictions for the number of blazars that will be resolved
380: by GLAST. Note that the Dermer luminosity functions were
381: \emph{not} used to determine the spectral shapes but solely to predict the numbers of BL Lacs and FSRQs GLAST will see, and thus, allowing us to estimate the reduction in the uncertainties on the
382: ISIDs. If more
383: blazars are, in fact, detected by GLAST, the ISIDs will be determined
384: with even greater confidence due to improved number statistics. Even
385: in the most conservative case, it is clear that GLAST observations
386: will place very tight constraints on the expected spectral shapes of
387: the unresolved emission of gamma-ray emitters - at least in the case
388: of blazar classes. Additionally, GLAST should also be able to further
389: constrain the EGRB itself. Thus, in light of GLAST, the spectral
390: shapes of the contributions of blazar populations would provide vital
391: information about whether those populations could, in principle,
392: explain all of the EGRB, or whether contributions from other gamma-ray
393: emitters are required.
394: %In minimizing the uncertainties in the expected spectral shapes, the
395: %%use of these shapes will become an effective tool in the effort to
396: %%%disentangle the possible contributions of different classes to the
397: %%%%EGRB.
398:
399: \section{Discussion}\label{Disc}
400:
401: We have calculated the expected spectral shape of the unresolved
402: gamma-ray emission from blazars under the assumptions that the ISIDs
403: of blazars do not evolve with redshift and are independent of blazar
404: luminosity and flaring state. We have also explicitly calculated the
405: $1\sigma$ uncertainty in the spectral shape entering through the
406: limited constraints on the blazar ISIDs derived from EGRET
407: data. Finally, we have predicted by how much these uncertainties will
408: be reduced if GLAST observations are used to determine the blazar
409: ISIDs.
410:
411: The unresolved emission spectral shape can be used as an indicator of
412: the potential importance of a given population's contribution to the
413: EGRB, and it constrains the maximal contribution at high energies
414: relative to that at low energies. If the curvature tentatively seen
415: in the observed EGRB is real, then a population with little such
416: curvature in its unresolved spectrum (such as the FSRQs) will not be
417: the dominant contributor to the EGRB at high energies even if it is
418: dominant at low energies. The unresolved BL Lac spectrum does seem to
419: be more convex than that of the FSRQs, but the level of uncertainty in the
420: spectral shape is high in this case because only a few BL Lacs were
421: detected by EGRET. However, GLAST observations will dramatically
422: improve our understanding of the blazar ISIDs and the associated
423: unresolved spectral shapes allowing us to use spectral shape
424: information to calculate the minimal additional contribution required
425: from other classes of sources to explain the observed EGRB spectrum.
426:
427: It should be stressed that here we have only calculated {\em
428: unabsorbed} spectra. At energies higher than 10 ${\rm \, GeV}$,
429: gamma-ray absorption through interactions (pair production) with the
430: extragalactic background light becomes important (e.g., Salamon \&
431: Stecker 1998), and the spectral shape of any contribution to the EGRB
432: will be accordingly changed. We will return to this effect in a future
433: publication.
434:
435: \acknowledgements{We gratefully acknowledge enlightening discussions
436: with Brian Fields, Julie McEnery, Angela Olinto, Martin Pohl, Jenny
437: Siegal-Gaskins, Floyd Stecker, and Kostas Tassis. This work was supported in part by the Kavli Institute for Cosmological Physics at the University of Chicago through grants NSF PHY-0114422 and NSF PHY-0551142 and an endowment from the Kavli Foundation and its founder Fred Kavli. T.M.V. was supported by an NSF Graduate
438: Research Fellowship.
439:
440: \begin{thebibliography}{}
441:
442: \bibitem[Brecher \& Burbidge(1972)]{1972ApJ...174..253B} Brecher, K., \&
443: Burbidge, G.~R.\ 1972, \apj, 174, 253
444:
445: \bibitem[de Boer et al.(2004)]{db05} de Boer, W., Sander, C., Gladyshev, A.~V.\ \& Kazakov, D.~I. 2005, A\&A 444, 51
446: \bibitem[Dermer(2006)]{D06} Dermer, C.~D.\ 2007, \apj, 659, 958
447:
448: \bibitem[Fossati et al.(1998)]{F98} Fossati, G., Maraschi, L., Celotti, A., Comastri, A., \& Ghisellini, G.\ 1998, \mnras, 299, 433
449:
450: \bibitem[Ghisellini et al.(1998)]{G98} Ghisellini, G., Celotti, A., Fossati, G., Maraschi, L., \& Comastri, A.\ 1998, \mnras, 301, 451
451:
452: \bibitem[Hartman et al.(1999)]{har99} Hartman, R.~C., et al.\
453: 1999, \apjs, 123, 79
454:
455: \bibitem[Lichti et al.(1978)]{lic78} Lichti, G.~G., Bignami, G.~F., \& Paul, J.~A.\ 1978, ApSS, 56, 403
456:
457: \bibitem[Maraschi et al.(2007)]{mar07} Maraschi, L., Ghisellini, G., \& Tavecchio, F.\ 2007, AIP Conference Proceedings, 921, 160
458:
459: \bibitem[M{\"u}cke \& Pohl(2000)]{mp00} M{\"u}cke, A., \&
460: Pohl, M.\ 2000, \mnras, 312, 177
461:
462: \bibitem[Mukherjee et al.(1997)]{Metal97} Mukherjee, R., et al. \ 1997,
463: \apj, 490, 116
464:
465: \bibitem[Mukherjee \& Chiang(1999)]{mc99} Mukherjee, R., \&
466: Chiang, J.\ 1999, Astroparticle Physics, 11, 213
467:
468: \bibitem[Narumoto \& Totani(2006)]{NT06} Narumoto, T., \&
469: Totani, T.\ 2006, \apj, 643, 81
470:
471: \bibitem[Padovani et al.(1993)]{pad93} Padovani, P.,
472: Ghisellini, G., Fabian, A.~C., \& Celotti, A.\ 1993, \mnras, 260, L21
473:
474: \bibitem[Padovani(2007)]{pad07} Padovani, P.\ 2007, \apss, 309, 63
475:
476: %\bibitem{pv07} Pavlidou, V. \& Venters, T.M. 2007, ApJ {\it in press}, arXiv:0%704.2417
477:
478: \bibitem[Pavlidou \& Fields(2002)]{pf02} Pavlidou, V., \&
479: Fields, B.~D.\ 2002, \apjl, 575, L5
480:
481: \bibitem[Pavlidou et al.(2007)]{petal07} Pavlidou, V., Siegal-Gaskins, J.~M, Fields, B.~D., Olinto, A.~V., \& Brown, C.\ 2007, {\it sumbitted to ApJ}
482: \bibitem[Pohl et al.(1997)]{p97} Pohl, M., Hartman, R.~C.,
483: Jones, B.~B., \& Sreekumar, P.\ 1997, \aap, 326, 51
484:
485: \bibitem[Salamon \& Stecker(1998)]{sal98} Salamon, M.~H., \&
486: Stecker, F.~W.\ 1998, \apj, 493, 547
487: % Absorption of High-Energy Gamma Rays
488:
489: \bibitem[Shen et al. (2006)]{she06} Shen, R., Kumar, P., \& Robinson, E.~L.\ 2006, MNRAS, 371, 1441.
490:
491: \bibitem[Stecker \& Salamon(1996)]{ste96a} Stecker, F.~W., \&
492: Salamon, M.~H.\ 1996a, \apj, 464, 600
493: % 1996 SS blazar model
494:
495: \bibitem[Stecker \& Salamon(1996)]{ste96b} Stecker, F.~W. \&
496: Salamon, M.~H.\ 1996b, Physical Review Letters, 76, 3878
497: % Rebuttal on Dar \& Shaviv for gamma-ray background from normal galaxies
498:
499: \bibitem[Strong et al.(2004)]{smr04} Strong, A.~W.,
500: Moskalenko, I.~V., \& Reimer, O.\ 2004, \apj, 613, 956
501:
502: \bibitem[Thompson et al.(1995)]{tetal95} Thompson, D.~J.\ et al.\ 1995, ApJS, 101, 259
503:
504: \bibitem[Venters \& Pavlidou (2007)]{vp07} Venters, T.~M. \& Pavlidou, V., ApJ {\it in press}, arXiv:0704.2417
505:
506: \end{thebibliography}
507:
508:
509:
510:
511: \end{document}
512: