1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass{emulateapj}
3:
4: \def\msun{M_{\odot}}
5: \def\simlt{\mathrel{\rlap{\lower 3pt\hbox{$\sim$}}\raise 2.0pt\hbox{$<$}}}
6: \def\simgt{\mathrel{\rlap{\lower 3pt\hbox{$\sim$}} \raise 2.0pt\hbox{$>$}}}
7: \def\lsim{\mathrel{\rlap{\lower 3pt\hbox{$\sim$}}\raise 2.0pt\hbox{$<$}}}
8: \def\gsim{\mathrel{\rlap{\lower 3pt\hbox{$\sim$}} \raise 2.0pt\hbox{$>$}}}
9: \def\di{\mbox{d}}
10: \def\Msun{{\rm M}_{\odot}}
11: \def\Zsun{{\rm Z}_{\odot}}
12:
13:
14: \shortauthors{Salvaterra \& Chincarini}
15: \shorttitle{The GRB Luminosity Function}
16:
17: \begin{document}
18:
19: \title{The Gamma Ray Burst Luminosity Function in the Light of the {\it Swift} 2--year Data}
20:
21: \author{
22: R.~Salvaterra,\altaffilmark{1}
23: G.~Chincarini,\altaffilmark{1,2}
24: }
25: \altaffiltext{1}{Dipertimento di Fisica G.~Occhialini, Universita degli Studi di Milano
26: Bicocca, Piazza della Scienza 3, I-20126 Milano, Italy, salvaterra@mib.infn.it}
27: \altaffiltext{2}{INAF, Osservatorio Astronomico di Brera, via E. Bianchi 46, I-23807 Merate (LC), Italy}
28:
29:
30: \begin{abstract}
31: We compute the luminosity function (LF) and the formation rate of long gamma
32: ray bursts (GRBs) by fitting the observed differential peak flux distribution
33: obtained by the {\it BATSE} satellite in three different
34: scenarios: i) GRBs follow the cosmic star formation and their LF is constant
35: in time; ii) GRBs follow the cosmic star formation but the LF varies with
36: redshift; iii) GRBs form preferentially in low--metallicity
37: environments. We find that the differential peak flux number counts
38: obtained by {\it BATSE} and by {\it Swift} can be reproduced using
39: the same LF and GRB formation rate, indicating that
40: the two satellites are observing the same GRB population.
41: We then check the resulting redshift distributions in the light of
42: {\it Swift} 2--year data, focusing in
43: particular on the relatively large sample of GRBs detected
44: at $z>2.5$. We show that models in which GRBs trace the cosmic star
45: formation and are described by a constant LF are ruled out by the number
46: of high--$z$ {\it Swift} detections. This conclusion does not depend on
47: the redshift distribution of bursts that lack
48: of optical identification, nor on the existence of a decline in star formation
49: rate at $z>2$, nor on the adopted faint--end of the GRB LF. {\it Swift}
50: observations can be explained by assuming that the LF varies with redshift
51: and/or that GRB formation is limited to low--metallicity environments.
52: \end{abstract}
53:
54: \keywords{gamma--ray: burst -- stars: formation -- cosmology: observations.}
55:
56: \section{Introduction}
57:
58: Gamma Ray Bursts (GRBs) are powerful flashes of high--energy photons occurring
59: at an average rate of a few per day throughout the universe. Even
60: though they are highly transient events very hard to localize, they are
61: so bright that they can be detected up to very high redshift (the current
62: record is $z=6.29$). The energy source of a GRB is believed
63: to be associated to the collapse of the core of a massive star in the case
64: of long--duration GRBs, and due to merger-- or accretion--induced collapse
65: for the short--hard class of GRBs (see M\'esz\'aros 2006 for a recent review).
66: In this paper, we limit our analysis to the class of long--duration GRBs.
67:
68: One of the main goals of the
69: {\it Swift} satellite (Gehrels et al. 2004) is to trackle the key issue
70: of the GRB luminosity function (LF). Unfortunately, although the number of
71: GRBs with good redshift
72: determination has been largely increased by {\it Swift},
73: the sample is still too poor (and bias dominated) to allow a direct
74: measurement of the LF.
75: Many studies (e.g. Lamb \& Reichart 2000; Porciani \& Madau 2001 (PM01); Schmidt
76: 2001, Choudhury \& Srianand 2002; Firmani et al. 2004; Guetta, Piran \& Waxman
77: 2005; Natarajan et al.2005; Daigne, Rossi \& Mochkovitch 2006)
78: tried to constrain the GRB LF under the assumption that GRBs trace the
79: observed star formation rate, as suggested by the association of long GRBs to
80: the death of massive stars.
81: Following these works and assuming the most
82: recent star formation rate determination, we
83: derive the LF and formation rate of GRBs by fitting the observed
84: {\it BATSE} differential peak flux distribution in three different
85: scenarios: i) GRBs
86: follow the cosmic star formation and have a constant LF; ii) the
87: GRB LF varies with redshift; iii) GRBs form in low--metallicity
88: environments. We check the results against the 2--year {\it Swift} data,
89: focusing in particular on the large sample of high redshift ($z>2.5$) GRBs
90: detected by this instrument.
91:
92: %The paper is organized as follow. In Section~2, we present the main equations
93: %we needed to compute number counts and redshift distribution. The different
94: %models explored are presented in Section~3, whereas the main results are
95: %discussed in Section~4 and 5. In Section 6 we estimate the detection rate of
96: %GRBs at $z\gsim 6$. Finally, in Section~7 we draw our conclusions.
97:
98:
99:
100:
101:
102: \section{Basic Equations}
103:
104: The observed photon flux, $P$, in the energy band
105: $E_{\rm min}<E<E_{\rm max}$, emitted by an isotropically radiating source
106: at redshift $z$ is
107:
108: \begin{equation}
109: P=\frac{(1+z)\int^{(1+z)E_{\rm max}}_{(1+z)E_{\rm min}} S(E) dE}{4\pi d_L^2(z)},
110: \end{equation}
111:
112: \noindent
113: where $S(E)$ is the differential rest--frame photon luminosity of the source,
114: and $d_L(z)$ is the luminosity distance.
115: To describe the typical burst spectrum we adopt the
116: functional form proposed by Band et al. (1993), i.e. a broken power--law
117: with a low--energy spectral index $\alpha$, a high--energy spectral index
118: $\beta$, and a break energy $E_b$. In this work, we take $\alpha=-1$ and
119: $\beta=-2.25$ (Preece et al. 2000), and $E_b=511$ keV (PM01).
120: Moreover, it is customary to define an isotropic equivalent intrinsic burst
121: luminosity in the energy band 30-2000 keV as
122: $L=\int^{2000\rm{keV}}_{30\rm{keV}} E S(E)dE$. Given a normalized GRB LF,
123: $\phi(L)$, and the detector efficiency, $\epsilon(P)$, the observed rate of
124: bursts with peak flux between $P_1$ and $P_2$ is
125:
126: \begin{eqnarray}
127: \frac{dN}{dt}(P_1<P<P_2)&=&\int_0^{\infty} dz \frac{dV(z)}{dz}
128: \frac{\Delta \Omega_s}{4\pi} \frac{\Psi_{\rm GRB}(z)}{1+z} \nonumber \\
129: & \times & \int^{L(P_2,z)}_{L(P_1,z)} dL^\prime \phi(L^\prime)\epsilon(P),
130: \end{eqnarray}
131:
132: \noindent
133: where $dV(z)/dz=4\pi c d_L^2(z)/[H(z)(1+z)^2]$ is the comoving volume
134: element\footnote{We adopted the 'concordance' model values for the
135: cosmological parameters: $h=0.7$, $\Omega_m=0.3$, and $\Omega_\Lambda=0.7$.},
136: and $H(z)=H_0 [\Omega_M (1+z)^3+\Omega_\Lambda+(1-\Omega_M-\Omega_\Lambda)(1+z)^2]^{1/2}$.
137: $\Delta \Omega_s$ is the solid angle covered on the sky by the survey,
138: and the factor $(1+z)^{-1}$ accounts for cosmological time dilation.
139: Finally, $\Psi_{\rm GRB}(z)$ is the comoving burst formation rate. In this
140: work, we assume that the GRB LF is described by
141:
142: \begin{equation}
143: \phi(L) \propto \left(\frac{L}{L_{\rm cut}}\right)^{-\xi} \exp \left(-\frac{L_{\rm cut}}{L}\right).
144: \end{equation}
145:
146:
147: \section{Models}
148:
149:
150: We consider three different scenarios. In the first one,
151: the GRB formation rate is proportional to the cosmic star formation rate
152: (SFR), $\Psi_\star(z)$, i.e. $\Psi_{\rm GRB}(z)=k_{\rm GRB} \Psi_\star (z)$,
153: and the LF does not evolve with redshift, i.e.
154: $L_{\rm cut}={\rm const}=L_0$. The factor $k_{GRB}$ gives the number of
155: GRBs formed per solar mass in stars and has units of $\Msun^{-1}$.
156: $\Psi_\star(z)$ (in units of $\Msun$ Mpc$^{-3}$ yr$^{-1}$) is
157: commonly parameterized with the form proposed by Cole et al. (2001) as
158:
159: \begin{equation}
160: \Psi_\star(z)=\frac{(a_1+a_2z)h}{1+(z/a_3)^{a_4}}.
161: \end{equation}
162:
163: \noindent
164: Recently, Hopkins \& Beacom (2006) have provided the values of the coefficients
165: $a$ by fitting the available UV and far--infrared measurements for $z<6$,
166: corrected for dust obscuration. In this paper, we adopt their best fit
167: parameters: $a_1=0.017$, $a_2=0.13$, and $a_3=3.3$ (Hopkins \& Beacom 2006).
168: The value of $a_4=4.3$ is
169: taken to be slightly lower than the original one in order to match the
170: decline of the SFR with $(1+z)^{-3.3}$ at $z\gsim 5$ suggested by recent
171: deep--field data (see Stark et al. 2006 and references therein).
172:
173: In the second scenario, while the GRB formation rate is still proportional to
174: the observed SFR, the cut--off luminosity in the GRB LF increases with
175: redshift as $L_{\rm cut}=L_0 (1+z)^\delta$. Lloyd--Ronning, Fryer \&
176: Ramirez--Ruiz (2002), using GRB redshifts and luminosities
177: derived from the luminosity--variability relationship, found that the data
178: imply $\delta\simeq 1.4\pm0.5$, and we adopt this as fiducial value.
179:
180:
181: Finally, we consider a case in which GRBs form only in environments with
182: metallicity below a given threshold (no evolution in the LF is considered).
183: In fact, some theoretical models (see M\'esz\'aros 2006 and reference therein)
184: require that GRB progenitors should have metallicity $\lsim 0.1\;\Zsun$.
185: Observations of GRB host galaxies (see Savaglio 2006 and
186: reference therein) seems in agreement with this prescription, showing that GRB
187: preferentially originates in low--metallicity regions.
188: Langer \& Norman (2006) have quantified the amount of star formation at a given
189: metallicity, using a recent determination of the stellar mass function
190: (Panter et al. 2004) and the observed mass--metallicity correlation
191: (Savaglio et al. 2005). Adopting the metallicity redshift evolution derived
192: from emission line studies (Kewley \& Kobulnicky 2005), the fractional mass
193: density belonging to metallicity below a given threshold, $Z_{th}$, can be
194: computed as
195:
196: \begin{equation}
197: \Sigma(z)=\frac{\hat{\Gamma}(0.84,(Z_{th}/\Zsun)^2 10^{0.3z})}{\Gamma (0.84)},
198: \end{equation}
199:
200: \noindent
201: where $\hat{\Gamma}$ ($\Gamma$) are the incomplete (complete) gamma
202: function, and $\Gamma(0.84)\simeq 1.122$. The GRB formation rate is then
203: given by $\Psi_{\rm GRB}(z)= k_{\rm GRB}\Sigma(z)\Psi_\star(z)$.
204: The main effect of this convolution is that the GRB
205: formation rate peaks at higher redshift with respect to the cosmic SFR.
206: We adopt $Z_{th}=0.1\;\Zsun$ as fiducial value, and, in this case, the GRB
207: formation peaks at $z\sim 3.5$.
208:
209:
210: \section{GRB number counts}
211:
212:
213: The free parameters in our model are the GRB formation efficiency
214: $k_{\rm GRB}$, the cut--off luminosity at $z=0$, $L_0$, and the power index,
215: $\xi$, of the GRB LF function. Following PM01, we optimized
216: the value of these parameters by $\chi^2$ minimization over the observed
217: differential number counts in the 50--300 keV band of {\it BATSE}.
218: We use the off--line
219: {\it BATSE} sample of Kommers et al. (2000), which includes 1998 archival
220: (``triggered'' plus ``non--triggered'') bursts, and for which the detector
221: efficiency is well described by the function
222: $\epsilon(P)=0.5[(1+{\rm erf}(-4.801+29.868 P)]$ (Kommers et al. 2000).
223: We report the best--fit parameters for our fiducial models in
224: Table~\ref{tab:fit}.
225: In the last column, we give the reduced $\chi^2$ for the
226: best--fitting model, showing that it is always possible to find a good
227: agreement with the data\footnote{Note that strong covariance on $L_0$ and
228: $\xi$ is observed in the parameter space surrounding the best--fit parameters
229: (see also PM01)}.
230: %\footnote{\bf Note that the model
231: %parameters show some degenerancy as discussed by PM01. Strong covariance
232: %on $L_0$ and $\xi$ is observed in the parameter space sorrounding the
233: %best--fit parameteres (see also PM01).}.
234: Note that for the metallicity evolution scenario a higher GRB formation
235: efficiency is required, since GRBs form only in a (small) fraction
236: of star forming galaxies.
237:
238:
239:
240: We can now use the best--fit parameters to compute the expected differential
241: peak flux distribution of GRBs in the 15--150 keV band of the Burst Alert
242: Telescope (BAT) instrument onboard of {\it Swift}. The results are plotted
243: in Figure~\ref{fig:swift} and
244: compared with the observed {\it Swift}/BAT data points.
245: All models show a good agreement with the data without the need of any change
246: of the GRB LF and formation efficiency, indicating that
247: {\it BATSE} and {\it Swift} are observing essentially the same
248: population of GRBs. This conclusion is rather insensitive to
249: $20$\% variations of the adopted GRB spectrum parameters, i.e. for
250: the large majority of burst spectra (Kaneko et al. 2006).
251:
252:
253:
254: \section{GRB redshift distribution}
255:
256: Our model allows us to compute the expected redshift distribution of GRBs
257: detected by {\it Swift}. We decide to avoid the comparison between model
258: results and
259: the overall observed distribution of bursts with known redshift, since
260: this procedure implicitly assumes that the observed sample of GRBs with
261: redshift determination is representative of all detected sources.
262: Moreover, important information are missed by this kind of analysis: for
263: example, that many bright GRBs are identified at high redshift.
264: So, we try to answer this simple question: {\it is the redshift
265: distribution consistent with the number of {\it Swift} detections at
266: $z>2.5$ and $z>3.5$?}
267:
268: The cumulative number of GRBs, identified during the two years of the
269: {\it Swift} mission at $z>2.5$ (left panel) and $z>3.5$ (right
270: panel), is plotted in Figure~\ref{fig:zgt} toghether with model predictions. Note
271: that {\it Swift} detections are to be considered as a strong lower limit,
272: since many high--$z$ bursts can be missed by optical follow--up searches.
273:
274: The model with no LF evolution clearly underestimates the number of high redshift
275: GRB detections at any photon flux and no bright GRBs are predicted for $z>3.5$.
276: We checked that variations of the shape of the SFR do not affect
277: this result: even assuming a constant SFR at $z\gsim 2$, the model predictions
278: do not change significantly. In fact, for relatively bright GRBs, the rapid
279: decline in the LF strongly hampers the detection of high redshift bursts.
280: Furthermore, our analysis does not depend on the faint--end of the LF:
281: increasing the population of faint GRBs would decrease the number of
282: high--$z$ detections, strengthening our conclusion. So, models
283: in which GRBs trace the cosmic SFR and are described by a
284: constant LF, are ruled out by the large sample of high--$z$ {\it Swift}
285: GRBs.
286:
287: The number of high--$z$ {\it Swift} identifications can be justified
288: assuming that the LF varies with redshift. In this case, high--$z$ GRBs
289: are typically brighter than low--$z$ ones, so that are much easely detected.
290: Assuming that the luminosity increases as $(1+z)^{1.4}$, we find
291: many sources at $z>2.5$, but the model is barely consistent with the number
292: of bright GRBs at $z>3.5$. Since some high--$z$ sources can be
293: missed by optical follow--up searches, an even stronger evolution might be
294: required to explain the data.
295:
296: Finally, we consider the possibility that GRB formation is restricted to
297: low--metallicity environments. In this case, the peak of the GRB
298: formation is shifted towards higher redshift, so that
299: the probability of high--$z$ detections increases.
300: Assuming $Z_{th}=0.1\;\Zsun$, {\it Swift} identification are exceeded
301: both at $z>2.5$ and $z>3.5$ without requiring any evolution in the LF.
302: Thus, the model is consistent with a fraction of high redshift bursts missed
303: by optical follow--up searches. Increasing the threshold metallicity will
304: decrease the number of sources at high--$z$: for $Z_{th}\sim 0.4\;\Zsun$ the
305: model becomes inconsistent with the number of observed GRBs at $z>3.5$.
306: Higher threshold values would require evolution of the GRB
307: luminosity and/or a more gentle decline of the SFR at high redshift.
308:
309: In conclusion, the existence of a large sample of bursts at
310: $z>2.5$ in the {\it Swift} 2-year data imply that
311: GRBs have experienced some kind of evolution, being more luminous or
312: more common in the past.
313:
314:
315: \section{GRB rate at redshift larger than six}
316:
317: The discovery of GRB050904 at $z=6.29$ (Antonelli et al. 2005;
318: Tagliaferri et al. 2005; Kawai et al. 2006) during the first year of the
319: {\it Swift} mission has strengthened
320: the idea that many GRBs should be observed out to very high redshift
321: (e.g. Natarajan et al. 2005; Bromm \& Loeb 2006; Daigne et al. 2006).
322: Unfortunately, no other source at $z\gsim 6$ has been detected in the second
323: year of observations.
324:
325:
326: In Figure~\ref{fig:zgt6}, we plot the {\it Swift} detection rate expected for
327: the three scenarios here considered. Models without evolution predict almost
328: no sources to be detected at very high redshift. If luminosity evolution
329: ($\delta=1.4$) is allowed, $\sim 2$ bursts/yr should lie above
330: $z\sim 6$ for $P>0.2$ ph cm$^{-2}$ s$^{-1}$, whereas, in the metallicity
331: evolution scenario ($Z_{th}=0.1\;\Zsun$), we expect $\sim 8$ GRBs/yr,
332: one or two being at $z\gsim 8$.
333:
334: The detection rate are found to decrease rapidly with increasing peak fluxes.
335: Indeed, it is interesting to note that GRB050904 was relatively
336: bright, being its observed photon flux $P=0.658$
337: ph cm$^{-2}$ s$^{-1}$. At this limit, only $\sim 1$ (2) bursts/yr
338: would be at $z\gsim 6$, if luminosity (metallicity) evolution is assumed.
339: Thus, the lack of very high redshift identification in the 2nd year of
340: the {\it Swift} mission might be due to
341: practical difficulties in the optical follow--up of faint GRBs.
342: In fact, no GRB with observed photon fluxes below $0.5$ ph cm$^{-2}$ s$^{-1}$
343: has UVOT detection and only in a couple of cases a reliable redshift
344: determination was possible.
345: So, the identification of just one burst at $z\gsim6$ in two years of
346: {\it Swift} mission is not very surprising. On the contrary, the discovery
347: of GRB050904 may suggest that the {\it Swift} follow--up procedure is
348: working very well, at least for relatively bright bursts.
349:
350:
351:
352:
353: \section{Conclusions}
354:
355: We have computed the luminosity function and the formation rate of long GRBs
356: by fitting the {\it BATSE} differential peak flux number counts in three
357: different scenarios: i) GRBs follow the cosmic star
358: formation and have a redshift--independent LF; ii) the GRB LF varies with
359: redshift; iii) GRBs are associated with star
360: formation in low--metallicity environments. In all cases, it is possible to
361: obtain a good fit to the data by adjusting the model free parameters.
362: Moreover, using the same LF and formation rate, it is possible to
363: reproduce both {\it BATSE} and {\it Swift} differential counts, showing
364: that the two satellites are observing the same GRB population.
365:
366: We have then computed the expected burst redshift distribution, testing the
367: results against the number of high redshift GRBs, detected during the two
368: years of the {\it Swift} mission. We find that models where GRBs trace the
369: SFR and are described by a constant LF largely underestimate the number
370: high--$z$ GRBs detected by {\it Swift}. This conclusion
371: does not depend on the redshift distribution of burst lacking
372: of optical identification, nor on the existence of a decline in the SFR at
373: $z>2$, nor on the adopted faint--end of the LF.
374: Alternatively, we find that the observed number of high--$z$
375: detection can be justified by assuming that the GRB luminosity increases with
376: redshift and/or that GRBs preferentially form in low--metallicity environments.
377:
378: Finally, we have estimated the detection rate of bursts at very high
379: redshift. We find that $\sim 2$ (8) GRBs/yr should be observed
380: at $z\gsim 6$, if luminosity (metallicity) evolution is assumed.
381: The majority of these sources is faint and may be missed in optical
382: follow--up searches, but $\sim 1$ (3) GRB/yr should be relatively
383: bright, with an observed photon flux in excees to 0.5 ph cm$^{-2}$ s$^{-1}$.
384:
385:
386: \begin{thebibliography}{}
387: \item{Antonelli, L. A., et al. 2005, GNC, 3924, 1}
388: \item{Band, D., et al. 1993, ApJ, 413, 281}
389: \item{Bromm, V., \& Loeb, A. 2006, ApJ, 642, 382}
390: \item{Choudhury, T. R., \& Srianand, R. 2002, MNRAS, 336, 27}
391: \item{Cole, S., et al. 2001, MNRAS, 326, 255}
392: \item{Daigne, F., Rossi, E. M., \& Mochkovitch, R. 2006, MNRAS, 372, 1034}
393: \item{Firmani, C. Avila--Reese, V., Ghisellini, G., \& Tutukov, A. V.
394: 2004, ApJ, 611, 1033}
395: \item{Gehrels, N., et al. 2004, ApJ, 611, 1005}
396: \item{Guetta, D., Piran, T., \& Waxman, E. 2005, ApJ, 619, 412}
397: \item{Hopkins, A. M., \& Beacom, J. F. 2006, ApJ, 651, 142}
398: \item{Kaneko, Y., Preece, R. D., Briggs, M. S., Paciesas, W. S., Meegan, C. A., \& Band, D. L., 2006, ApJS, 166, 298}
399: \item{Kawai, N., et al. 2006, Nature, 440, 184}
400: \item{Kewley, L., \& Kobulnicky, H.A. 2005, in: {\it Starburst: From 30
401: Doradus to Lyman Break Galaxies}, R. de Grijs and R. M. Gonzalez Delgado,
402: Astrophysics \& Space Science Library, Vol. 329. Dordrecht: Springer, p. 307}
403: \item{Kommers, J. M., Lewin, W. H. G., Kouveliotou, C., van Paradijs, J.,
404: Pendleton, G. N., Meegan, C. A., \& Fishman, G. J. 2000, ApJ, 533, 696}
405: \item{Lamb, D. Q., \& Reichart, D. E. 2000, ApJ, 536, 1}
406: \item{Langer, L., \& Norman, C. A. 2006, ApJ, 638, L63}
407: \item{Lloys--Ronning, N. M., Fryer, C. L., \& Ramirez--Ruiz, E. 2002, ApJ,
408: 574, 554}
409: \item{M\'esz\'aros, P. 2006, Reports of Progress in Physics, 69, 2259}
410: \item{Natarajan, P., Albanna, B., Hjorth, J., Ramirez--Ruiz, E., Tarvir, N.,
411: \& Wijers, R. 2005, MNRAS, 364, L8}
412: \item{Panter, B., Heavens, A. F., \& Jimenez, R. 2004, MNRAS, 355, 764}
413: \item{Porciani, C., \& Madau, P. 2001, ApJ, 548, 522 (PM01)}
414: \item{Preece, R. D., Briggs, M. S., Mallozzi, R. S., Pendleton, G. N.,
415: Paciesas, W. S., \& Band, D. L. 2000, ApJS, 126, 19}
416: \item{Savaglio, S., et al. 2005, ApJ, 635, 260}
417: \item{Savaglio, S. 2006, NJPh, 8, 195}
418: \item{Schmidt, M. 2001, ApJ, 552, 36}
419: \item{Stark, D. P., Bunker, A. J., Ellis, R. S., Eyles, L. P.,\& Lacy, M.
420: 2006, ApJ in press, astro-ph/0604250}
421: \item{Tagliaferri, G., et al. 2005, A\&A, 443, L1}
422: \end{thebibliography}
423:
424: \clearpage
425:
426: %TABLE
427: \begin{deluxetable}{lcccc}
428: \tablecolumns{5}
429: \tablewidth{0pt}
430: \tablecaption{Best fit parameters for different models. Errors are at 1$\sigma$
431: level.}
432: \tablehead{\colhead{Model} & \colhead{$k_{\rm GRB}/(10^{-8}\Msun^{-1})$} & \colhead{$L_0/(10^{51} {\rm \;erg\;s}^{-1})$} & \colhead{$\xi$} & \colhead{$\chi_r^2$}}
433: \startdata
434: no evolution & $1.14\pm0.07$ & $9.54\pm4.55$ & $3.54\pm0.78$ & 0.83 \\
435: luminosity evolution ($\delta=1.4$) & $1.05\pm0.05$ & $0.77\pm0.13$ & $2.19\pm0.95$ & 0.80 \\
436: metallicity evolution ($Z_{th}=0.1\;\Zsun$) & $10.0\pm0.5$ & $16.7\pm5.7$ & $2.94\pm0.34$ & 0.84 \\
437: \enddata
438: \label{tab:fit}
439: \end{deluxetable}
440:
441:
442: \begin{figure}
443: \epsscale{1.0}
444: \plotone{f1.ps}
445: \caption{Differential number counts for {\it Swift} in the 15--150 keV band
446: as a function of the observed photon flux $P$. The points show the observed
447: counts and their Poisson uncertainties (horizontal error bars denote bin
448: size). Dotted lines refers to the model without evolution, short--dashed line
449: to the luminosity evolution model ($\delta=1.4$), and
450: long--dashed line to the model with the metallicity threshold for GRB formation
451: ($Z_{th}=0.1\;\Zsun$). A field of view of 1.4 sr for {\it Swift}/BAT is
452: adopted.}
453: \label{fig:swift}
454: \end{figure}
455:
456: \clearpage
457:
458: \begin{figure}
459: \epsscale{1.0}
460: \plotone{f2.ps}
461: \caption{Cumulative number of high redshift GRBs at $z>2.5$ (right panel) and
462: at $z>3.5$ (left panel) as a function of the observed photon flux $P$ in the
463: 15--150 keV band. The number of sources detected in the two years of
464: {\it Swift} mission is shown as solid histogram, whereas model results are
465: shown with lines as in the previous figure. Note that the observed detections
466: are lower limits, since many high--$z$ GRBs can be missed by optical
467: follow--up searches. A field of view of 1.4 sr for {\it Swift}/BAT is
468: adopted.}
469: \label{fig:zgt}
470: \end{figure}
471:
472: \clearpage
473:
474: \begin{figure}
475: \epsscale{1.0}
476: \plotone{f3.ps}
477: \caption{Cumulative rate of $z\gsim 6$ GRBs detectable by {\it Swift} as a
478: function of the photon flux $P$. A field of view of 1.4 sr for
479: {\it Swift}/BAT is adopted. Lines as in Figure~\ref{fig:swift}.}
480: \label{fig:zgt6}
481: \end{figure}
482:
483: \end{document}
484: