1: \documentclass[12pt,preprint]{aastex}
2:
3: \begin{document}
4:
5: \title{A NEW DISTANCE ESTIMATOR AND
6: SPATIAL DISTRIBUTION OF GRBS OBSERVED BY BATSE}
7:
8: \author{Heon-Young Chang$^1$, Suk-Jin Yoon$^2$, and Chul-Sung Choi$^3$}
9:
10: \affil{$^1$
11: Korea Institute for Advanced Study\\
12: 207-43 Cheongryangri-dong Dongdaemun-gu, Seoul 130-012, Korea,\\
13: $^2$
14: Center for Space Astrophysics and Department of Astronomy,
15: Yonsei University\\
16: 134 Shinchon-dong Seodaemun-gu, Seoul 120-749, Korea\\
17: $^3$
18: Korea Astronomy Observatory \\
19: 36-1 Hwaam-dong, Yusong-gu, Taejon 305-348, Korea}
20: \email{hyc@ns.kias.re.kr, sjyoon@csa.yonsei.ac.kr, cschoi@kao.re.kr}
21:
22: \begin{abstract}
23: We propose a redshift estimator for the long ($T_{90} > 20$)
24: gamma-ray bursts (GRBs) observed by the BATSE.
25: It is based on an empirical relation between
26: the redshift and the power-law index of power density spectra (PDSs)
27: of the observed GRBs. This relation is constructed by using the fact that
28: the power-law index is dependent upon
29: a characteristic timescale of GRB light curves which are
30: inevitably stretched by cosmological time dilation.
31: We construct the empirical relation using
32: both individual PDSs and averaged PDSs.
33: An error estimates of $z$ are 1.09 and 1.11 for the empirical relation
34: by individual PDS fits,
35: 1.72 and 1.56 by averaged PDS fits,
36: for the least squares fit and the maximum likelihood fit, respectively.
37: We attempt to determine the spatial
38: distribution of the GRBs observed by the BATSE as a function of
39: redshifts on the basis of the resulting redshift estimator.
40: We find that the obtained spatial distribution of the
41: observed GRBs seems consistent with that of the GRBs whose redshifts
42: are reported, even though the estimated errors are not very accurate.
43: The GRBs observed by the BATSE seem distributed within $z \sim 5-6$.
44: This result has implications on theoretical calculations
45: of stellar formations at high redshifts and beaming geometry via
46: a statistical study of the observed GRBs involving
47: beaming-induced luminosity functions.
48: We discuss such implications, and possible
49: uncertainties of the suggested method.
50: \end{abstract}
51: \keywords{cosmology:theory --- gamma rays:bursts --- methods:data analysis}
52:
53: \section{Introduction}
54:
55: Since gamma-ray bursts (GRBs) were first discovered in late 60's
56: (Klebesadel et al. 1973),
57: thousands of GRBs have been detected up to date.
58: The discovery of afterglows
59: in other spectral bands and host galaxies enabled us to
60: measure redshifts of about twenty GRBs
61: (see, e.g., ${\rm http://www.aip.de/\sim jcg/grbgen.html}$),
62: establishing the fact that GRBs
63: are indeed cosmological
64: (Mao and Paczy${\rm \acute{n}}$ski 1992; Meegan et al. 1992;
65: Piran 1992; Metzger et al. 1997).
66: Nonetheless, the redshifts are still unknown
67: for most of the detected GRBs. Unless we locate
68: a burst on the sky by immediate follow-up observations, the distance
69: of the burst is apt to remain unrevealed forever.
70:
71: Concerning physical models of GRBs,
72: their distance scales are related to key issues,
73: such as, energetics, burst rates as a function of redshifts.
74: That is, estimating the distance puts direct constraints
75: on the theories of the observed GRBs.
76: Besides this
77: the redshift distribution of GRBs should track the cosmic star
78: formation rate of massive stars,
79: if GRBs are indeed related to the collapse of massive stars
80: (Woosley 1993; Paczy${\rm \acute{n}}$ski 1998; MacFadyen and Woosley 1999).
81: Therefore, once its association has been proven, one expects
82: the observed GRBs are the most powerful probe of the high redshift universe
83: (Wijers et al. 1998; Blain and Natarajan 2000; Lamb and Reichart 2000).
84: In fact, the GRB formation rate and the star formation rate (SFR)
85: have similar slopes at low redshift,
86: implying that GRBs can be used indeed
87: as a probe of the cosmic star formation rate at high redshift.
88: Observations of faint galaxies have been used to estimate the history of
89: star formation activity (Madau et al. 1996; Rowan-Robinson 1999;
90: Steidel et al. 1999). However, there are considerable difficulties and
91: uncertainties in the corrections that should be applied due to
92: extinction and obscuration. An independent determination and
93: test of the relative amount of obscured and unobscured star formation
94: activity would be extremely valuable.
95:
96: At present, there are too few redshift
97: measurements with which to produce the global GRB formation rate.
98: This fact is indeed hard to avoid unless observers set up networks
99: of efficient telescopes in order for an immediate follow-up observations.
100: Recently, however, there are pilot studies
101: to overcome the technical ability mentioned above
102: (e.g., Reichart et al. 2001), though
103: there have been several attempts to quantify pulse shapes of GRBs
104: and interpret results in terms of GRB physics
105: (Fenimore et al. 1996; Norris et al. 1996; In'T Zand and Fenimore 1996;
106: Kobayashi et al. 1997; Daigne and Mochkovitch 1998; Fenimore 1999;
107: Panaitescu et al. 1999).
108: Several authors (Stern et al. 1999; Fenimore and Ramires-Ruiz 2000;
109: Reichart et al. 2001)
110: began to observe strong correlations between temporal
111: properties of the observed GRBs and their brightness, which may
112: have some implications that the measured
113: spikiness can be used to obtain
114: distances much like a Cepheid-like distance estimator.
115: Norris et al. (2000) also showed the spectral lag/luminosity relationship for
116: six bursts with known redshifts can be appreciated.
117: Currently, the luminosity estimator yields best-estimate
118: luminosity distances that are accurate to a factor
119: of $\approx 2$ (see Reichart et al. 2001).
120:
121: Along the line of efforts of such kinds
122: we propose a new method based on an empirical relation motivated
123: by the work of Chang (2001).
124: It is well known that power density spectra
125: (PDSs) of long GRBs show a power-law behavior (Beloborodov et al. 1998, 2000).
126: Though its underlying physical mechanism is not obvious
127: (Panaitescu et al. 1999; Chang and Yi 2000),
128: the PDS analysis may provide useful information of physics of GRBs
129: (Panaitescu et al. 1999) and the distance information (Chang 2001).
130: Particularly, Chang (2001) has demonstrated that the power-law
131: index of PDSs of the observed GRBs shows a redshift dependence,
132: implying a possible relationship between the power-law
133: index and the redshift of GRBs.
134: It can be possibly worked out because of the fact that
135: burst profiles should be stretched in time due to
136: cosmological time dilation by an amount proportional to the redshift, $1+z$.
137:
138:
139: In \S 2 we begin with a brief summary of the PDS analysis in GRB studies,
140: and describe the empirical relation involved in our procedure.
141: In \S 3 we present results obtained by
142: applying our method to the GRBs observed by the BATSE instrument aboard
143: the {\it Compton Gamma Ray Observatory} (Paciesas et al. 1999),
144: and discuss what they suggest.
145: Finally, we conclude by pointing out that the accuracy of our
146: redshift estimates is limited by unknown underlying properties of
147: GRBs and what should be further developed in \S 4.
148:
149: \section{Empirical Relation of Power-law Index and Redshift}
150:
151: Contrary to the diverse and stochastic behavior in the time domain,
152: long GRBs show a simple behavior in the frequency domain
153: (e.g., Beloborodov et al. 1998). The power-law
154: behavior is seen even in a single burst when it is bright and long.
155: The power-law PDS provides a new tool for studies of GRBs
156: themselves. Using the PDS analysis, Panaitescu et al. (1999) analyzed
157: the temporal behavior of GRBs in the framework of a relativistic
158: internal shock model. They set up
159: their internal shock model and attempted to identify the most sensitive model
160: parameters to the observed PDS,
161: which is defined by the square of
162: the Fourier transform of the observed light curve.
163: They concluded that
164: the wind must be modulated such that collisions at large radii release
165: more energy than those at small radii in order to reproduce consistent
166: PDSs with the observation. However, it is also noted that
167: the reported power-law behavior with the index of $-5/3$ and
168: the exponential distribution of the observed PDS
169: can be reproduced by adjusting the sampling interval in the time domain
170: for a given decaying timescale of individual
171: pulse in a specific form of GRB light curves (Chang and Yi 2000).
172: Therefore, conclusions on the central engine of GRBs on the basis of
173: the PDS analysis should be derived with due care.
174:
175: Another valuable use of the PDS analysis can be realized bearing in mind
176: that for a given sampling interval
177: the resulting power-law index is dependent upon
178: the characteristic timescale of the observed light curve. For instance,
179: consider a GRB light curve as a sum of exponential functions of time, $f(t)$,
180: as considered in Chang and Yi (2000). Since the
181: Fourier transform of $f(at)$ is given by $\frac{1}{|a|}F(\nu /a)$, where
182: $F(\nu)$ is the Fourier transform of $f(t)$,
183: $\nu$ being the cyclic frequency and $a$ being
184: a constant, fitting of the power law function to the PDS
185: may result in a different power-law index
186: as the constant $a$ varies when the sampling interval is pre-determined.
187: In other words, for the observed GRB light curves with the pre-determined
188: sampling intervals, e.g., 64 ms, the cosmological time dilation stretches
189: the light curve by an amount of $1+z$
190: and consequently results in changes in the
191: obtained power-law index. This should be true because cosmological objects
192: like GRBs should not only be redshifted in energy but also extended
193: in time because of the expansion of the universe.
194: Chang (2001) demonstrated that a cosmological time dilation effect is
195: indeed imprinted in the light curves of the observed GRBs
196: whose redshifts are known by dividing
197: the GRBs into near and far groups. The author has showed that the near
198: GRB group ends up with the smaller
199: power-law index than the far one and that
200: the correction with the $1+z$ factor removes the differences.
201: The power-law index
202: difference in two separate groups is larger than
203: that among different energy bands.
204:
205: In order to construct the empirical relation between the power-law index
206: and the redshift of GRBs we have calculated the power-law index of the PDSs
207: of 9 GRBs detected by the BATSE with known redshifts.
208: We have used light curves of the GRBs from the
209: updated BATSE 64 ms ASCII
210: database\footnote{ ${\rm ftp://cossc.gsfc.nasa.gov/pub/data/batse/}$}. From this
211: archive we select the light curves of the GRBs in channel 2
212: whose redshifts are available.
213: We list up the GRBs used in our analysis with
214: their reported redshifts in Table 1.
215: We calculate
216: the Fourier transform of each light curve of the GRBs
217: and the corresponding PDS.
218: Before taking the Fourier transform of
219: light curves we scale them such that the height of their highest peak has
220: unity in the GRB light curves.
221: Since the individual PDS of GRBs are stochastic,
222: different parts of the PDS appear to follow a slightly different power-law index.
223: Having calculated the PDS of an individual GRB we obtain the power-law index
224: of the PDS using the limited part of the PDS, i.e., $-1.6 < \log \nu < 0$.
225: The lower bound is roughly determined in that the deviation from the
226: power law begins due to the finite length of bursts.
227: The upper bound is where the Poisson noise
228: becomes dominant.
229: Poisson noise in the measured count rate affects the PDS at high frequencies
230: and has a flat spectrum.
231: The Poisson noise level equals the burst total fluence including the
232: background in the considered time window. We calculate the individual
233: Poisson level for each burst and subtract it from the burst PDS.
234:
235: The PDSs can be described as a single power law with superimposed
236: fluctuations which follow the exponential distribution,
237: which may require the maximum likelihood method. However, by considering
238: that we smooth the PDSs on the scale $\Delta \log \nu = 0.5$ before fitting,
239: the least squares fit can be preferred since the error distribution
240: may be modified to the normal distribution according to the
241: {\it central limit theorem}. We use two different fitting
242: routines corresponding to the normal error distribution and the
243: exponential error distribution.
244: We employ both fitting algorithms to compare.
245: What is shown in Figure 1 are empirical relations of the redshift
246: and the power-law index obtained by the least squares fit
247: and the maximum likelihood method.
248: Firstly, we attempt to construct the empirical relation of the
249: redshift and the power-law index without averaging of the PDSs.
250: That is, we construct the best
251: fit using power-law indices of 9 individual data points.
252: Results are also shown in Figure 1, where
253: the thin solid line results from the least squares fit, the thin dashed
254: line the maximum likelihood method.
255: A possible alternative way to extract the empirical relation
256: from the noisy individual PDSs is to take the average PDSs over a sample
257: of long GRBs. Then the random fluctuations affecting each individual PDS tend
258: to cancel each other and the power-law behavior can be clearly seen.
259: Because of the small number of data, we group the PDSs into 4 subgroups
260: according to the reported redshifts, and average the PDSs before a fitting
261: process :
262: GRB 980329 + GRB 971214, GRB 990123 + GRB 990506 + GRB 990510, GRB 970508
263: + GRB 980703 + GRB 991216, GRB 980425.
264: The thick solid line results from fitting of averaged PDSs by
265: the least square fit, the thick dashed
266: line the maximum likelihood method.
267: We note that individual fitting results in a less steeper
268: relation. We also have attempted higher order polynomial fits but it did not
269: end up with a monotonic relation as one should expect.
270:
271: The error estimates of $z$, which are defined by a square root of
272: the average of squared difference between
273: the measured redshifts summarized in Table 1
274: and the expected redshifts by the fitting,
275: are 1.09 and 1.11 for the empirical relation due to individual PDS fits,
276: 1.72 and 1.56 due to the averaged PDS fits,
277: for the least squares fit and the maximum likelihood fit, respectively.
278:
279: \section{Spatial Distribution of GRBs}
280:
281: We adopt light curves of the long GRBs from the updated BATSE 64 ms ASCII
282: database as in processes above.
283: We choose bursts with durations $T_{90} > 20 $ s, where $T_{90}$ is
284: the time it takes to accumulate from 5 \% to 90 \% of the total fluence
285: of a burst summed over all the four channels. Of those bursts,
286: we further select bursts with the
287: peak count rates satisfying $C_{\rm max}/C_{\rm min}$
288: for the 64 ms trigger timescale is greater than 1.
289: Applying these criteria, we end up with 388 bursts.
290:
291: In a similar way, we obtain power-law indices of PDSs of the selected
292: GRBs and subsequently estimates of their redshifts.
293: In Figure 2, the redshift distributions of the GRBs obtained by
294: the relation we have in the previous section are shown. Different line types
295: indicate same meanings as in Figure 1. Note that the thin dotted histogram
296: represents the spatial distributions of the 22 GRBs whose redshifts are
297: available at the web site, from where the quoted redshifts in Table 1 are taken.
298: It is interesting to note that the predicted redshift distributions of
299: GRBs that are derived from fitting of individual power-law index
300: without averaging PDSs appear to provide a better agreement with
301: the redshift distribution of the GRBs with redshift-known.
302:
303: The obtained GRB redshift distribution is quite suggestive. Firstly,
304: it is seen in Figure 2 that the long GRBs observed by the BATSE, at least,
305: are distributed well within $z \sim 5-6$.
306: If one accepts an idea that the GRB formation rate should trace
307: the SFR at high redshifts as at low redshifts, this is in apparent
308: contrast to what is derived from some theoretical calculations of star formation.
309: Theoretical calculations show that the birth rate of Pop III stars produces
310: a peak in the SFR in the universe at redshifts $16 \la z \la 20$,
311: while the birth rate of Pop II stars produces a much larger and broader
312: peak at redshifts $2 \la z \la 10$ (Ostriker and Gnedin 1996). Secondly,
313: according to the cumulative redshift distributions derived in terms of
314: beaming-induced luminosity functions an extreme shape of a conic
315: beam seems likely to be ruled out : although the observed $<V/V_{\rm max}>$
316: can be satisfied with the theoretical $<V/V_{\rm max}>$,
317: a broad beam cannot explain the observed redshifts greater than $\sim 2 - 3$
318: (Kim et al. 2001) and a hollow beam expects too many GRBs farther than
319: $z \sim 5-6$ (Chang and Yi 2001), if a SFR-motivated
320: number density distribution of GRB sources is assumed.
321:
322: \section{Discussions}
323:
324: There are problems and limits to determine redshifts accurately
325: in both obtaining the relation and applying this relation to data.
326: First of all, even though this method is in principle to
327: work, it is not clear whether we may apply this method over the observed
328: GRB light curves obtained by various satellite missions at the same time.
329: It is partly because the PDS of each individual burst is composed of
330: the power law and superimposed exponentially distributed fluctuations
331: which make it difficult to recognize the power law in an individual burst,
332: and partly because such the empirical relation is susceptible to
333: observational conditions, such as, trigger timescale, detection sensitivity.
334: Availability of more redshifts of GRBs may help to reproduce a better
335: relation.
336: With all the efforts in implementing a sophisticated algorithm to accommodate
337: the diversity of the light curves, it is essential to understand
338: a fundamental mechanism of GRBs to derive
339: the intrinsic relation of the power-law index and the redshift.
340: Secondly, We have implicitly assumed that
341: all the long GRBs have a more or less
342: same characteristic timescale and cosmological time dilation alone
343: affects varying the characteristic timescale.
344: We need to understand clearly what and how forms the flat part of PDS.
345: Thirdly, the effect of redshift tends to flatten PDSs of GRBs
346: on the contrary to the time dilation effect.
347: It reflects a well-known fact that pulses in
348: a single GRB are more narrow in a higher energy band
349: (e.g., Norris et al. 1996).
350: These effects combine and produce undesirable results in obtained
351: power-law index. We have presumed in this study
352: that the time dilation effect on the power-law index is larger than
353: that of the redshift as observed in Chang (2001)
354: and ignored the effect of the redshift.
355: However, it should be understood how
356: the power-law index relates with the energy channels
357: to improve the proposed method accommodating the redshift effect.
358:
359: We conclude by pointing out that, even though redshift estimates
360: are subject to the stochastic nature of the observed PDSs and
361: accuracy of estimates are limited by unknown properties of the GRBs
362: the encouraging conclusion of this study is
363: that redshifts of the GRBs can be obtained with the GRB light curves,
364: whose redshifts otherwise remain unknown forever.
365:
366: \acknowledgements
367:
368: SJY is supported by the Creative Research Initiatives Program of
369: the Korean Ministry of Science and Technology.
370: This research has made use of data obtained through
371: the High Energy Astrophysics Science Archive
372: Research Center Online Service,
373: provided by the NASA/Goddard Space Flight Center.
374:
375: \newpage
376: \begin{thebibliography}{}
377: \bibitem[]{}
378: Blain, A. W. and Natarajan, P. 2000, MNRAS, 312, L35
379: \bibitem[]{}
380: Beloborodov, A., Stern, B., and Svensson, R. 1998, ApJ, 508, L25
381: \bibitem[]{}
382: Beloborodov, A., Stern, B., and Svensson, R. 2000, ApJ, 535, 158
383: \bibitem[]{}
384: Chang, H.-Y. 2001, ApJ, 557, L85
385: \bibitem[]{}
386: Chang, H.-Y. and Yi, I. 2001, ApJ, 554, 12
387: \bibitem[]{}
388: Chang, H.-Y. and Yi, I. 2000, ApJ, 542, L17
389: \bibitem[]{}
390: Daigne, F. and Mochkovitch, R. T. 1998, MNRAS, 296, 275
391: \bibitem[]{}
392: Fenimore, E. 1999, ApJ, 518, 375
393: \bibitem[]{}
394: Fenimore, E. E. and Ramirez-Ruiz, E. 2000, astro-ph/0004176
395: \bibitem[]{}
396: Fenimore, E., Madras, C. D., and Nayakshin, S. 1996, ApJ, 473, 998
397: \bibitem[]{}
398: In'T Zand, J. J. M. and Fenimore, E. 1996, ApJ, 464, 662
399: \bibitem[]{}
400: Kim, C., Chang, H.-Y., and Yi, I. 2001, ApJ, 548, 532
401: \bibitem[]{}
402: Klebesadel, R. W., Strong, I. B., and Olson, R. A., 1973, ApJ 182, L85
403: \bibitem[]{}
404: Kobayashi, S., Piran, T., and Sari, R. 1997, ApJ, 490, 92
405: \bibitem[]{}
406: Lamb, D. Q. and Reichart, D. E. 2000, ApJ, 536, 1
407: \bibitem[]{}
408: MacFadyen, A. I. and Woosley, S. E., 1999, ApJ, 524, 262
409: \bibitem[]{}
410: Madau, P., et al. 1996, MNRAS, 283, 1388
411: \bibitem[]{}
412: Mao, S. and Paczy${\rm \acute{n}}$ski, B. 1992, ApJ, 388, L45
413: \bibitem[]{}
414: Meegan, C. A., et al. 1992, Nature, 355, 143
415: \bibitem[]{}
416: Metzger, M. R., et al. 1997, Nature, 387, 879
417: \bibitem[]{}
418: Norris, J. P., et al. 1996, ApJ, 459, 393
419: \bibitem[]{}
420: Norris, J. P., Marani, G. F., and Bonnell, J. T. 2000, ApJ, 534, 248
421: \bibitem[]{}
422: Ostriker, J. P. and Gnedin, N. Y. 1996, ApJ, 472, L63
423: \bibitem[]{}
424: Paciesas, W. S., et al. 1999, ApJS, 122, 465
425: \bibitem[]{}
426: Paczy${\rm \acute{n}}$ski, B. 1998, ApJ, 494, L45
427: \bibitem[]{}
428: Panaitescu, A., Spada, M., and ${\rm M\acute{e}sz\acute{a}ros }$,
429: P. 1999, ApJ, 522, L105
430: \bibitem[]{}
431: Piran, T. 1992, ApJ, 389, L45
432: \bibitem[]{}
433: Reichart, D. E., et al. 2001, ApJ, 552, 57
434: \bibitem[]{}
435: Rowan-Robinson, M. 1999, Ap\&SS, 266, 291
436: \bibitem[]{}
437: Steidel, C. C., et al. 1999, ApJ, 519, 1
438: \bibitem[]{}
439: Stern, B., Poutanen, J., and Svensson, R. 1999, ApJ, 510, 312
440: \bibitem[]{}
441: Wijers, R. A. M. J., et al. 1998, MNRAS, 294, L13
442: \bibitem[]{}
443: Woosley, S. E. 1993, ApJ, 405, 273
444: \end{thebibliography}
445:
446: \newpage
447: %Figure 1
448: \begin{figure}
449: \plotone{f1.eps}
450: \caption{
451: Relation of power-law index and redshift.
452: The thin solid line results from fitting of individual PDSs
453: by the least squares fit, the thin dashed
454: line the maximum likelihood method.
455: For comparison, the relation of
456: redshift and power-law index with averaged PDSs is also shown.
457: The thick solid line and the thick dashed line represent
458: the least squares fit, the maximum likelihood method, respectively.
459: \label{fig1}}
460: \end{figure}
461:
462: %Figure 2
463: \begin{figure}
464: \plotone{f2.eps}
465: \caption{
466: Normalized
467: spatial distributions of 388 long GRBs observed by the BATSE are compared
468: with that of the redshift-known GRBs, indicated by the same line types
469: as in Figure 1 and the thin dotted line, respectively.
470: \label{fig2}}
471: \end{figure}
472:
473: \newpage
474: \begin{table}
475: \begin{center}
476: \caption{A list of the GRBs used in the analysis with
477: the redshifts and $T_{90}$.
478: The redshifts are quoted from a complied table at
479: ${\rm http://www.aip.de/\sim jcg/grbgen.html}$. \label{tbl-1}}
480: \vspace{2mm}
481: \begin{tabular}{cccc}
482: \hline
483: GRB name&trigger number&redshift&$T_{90}$ (secs) \\ \hline\hline
484: GRB 991216 & 7906 & 1.02 & 15 \\
485: GRB 990510 & 7560 & 1.619 & 68 \\
486: GRB 990506 & 7549 & 1.3 & 130 \\
487: GRB 990123 & 7343 & 1.60 & 63 \\
488: GRB 980703 & 6891 & 0.966 & 411 \\
489: GRB 980425 & 6707 & 0.0085 & 34 \\
490: GRB 980329 & 6665 & 3.9 & 18 \\
491: GRB 971214 & 6533 & 3.42 & 31 \\
492: GRB 970508 & 6225 & 0.835 & 23 \\ \hline
493: \end{tabular}
494: \end{center}
495: \end{table}
496:
497: \end{document}
498:
499: