1:
2: %% The command below calls the preprint style
3: %% which will produce a one-column, single-spaced document.
4: %% Examples of commands for other substyles follow. Use
5: %% whichever is most appropriate for your purposes.
6: %%
7: %%\documentclass[12pt,preprint]{aastex}
8:
9: %% manuscript produces a one-column, double-spaced document:
10:
11: %%\documentclass[manuscript]{aastex}
12:
13: %% preprint2 produces a double-column, single-spaced document:
14:
15: \documentclass[twocolumn]{emulateapj}
16: \usepackage{epstopdf}
17: %\documentclass[preprint2,emulateapj5]{aastex}
18:
19: %% Sometimes a paper's abstract is too long to fit on the
20: %% title page in preprint2 mode. When that is the case,
21: %% use the longabstract style option.
22:
23: %%\documentclass[preprint2,longabstract]{aastex}
24:
25: \newcommand{\myemail}{Wystan.Benbow@mpi-hd.mpg.de}
26: \newcommand{\hisemail}{Berrie.Giebels@poly.in2p3.fr}
27: \newcommand{\MPIK}{Max-Planck-Institut f\"ur Kernphysik, Heidelberg, Germany}
28: \newcommand{\Yerevan}{Yerevan Physics Institute, Armenia}
29: \newcommand{\Tolouse}{Centre d'Etude Spatiale des Rayonnements, CNRS/UPS, Toulouse, France}
30: \newcommand{\Hamburg}{Universit\"at Hamburg, Institut f\"ur Experimentalphysik, Germany}
31: \newcommand{\Berlin}{Institut f\"ur Physik, Humboldt-Universit\"at zu Berlin, Germany}
32: \newcommand{\Meudon}{LUTH, UMR 8102 du CNRS, Observatoire de Paris, Section de Meudon, France}
33: \newcommand{\Durham}{University of Durham, Department of Physics, U.K.}
34: \newcommand{\Potch}{Unit for Space Physics, North-West University, Potchefstroom, South Africa}
35: \newcommand{\LLR}{Laboratoire Leprince-Ringuet, IN2P3/CNRS,
36: Ecole Polytechnique, Palaiseau, France}
37: \newcommand{\Annecy}{Laboratoire d'Annecy-le-Vieux de Physique des Particules, IN2P3/CNRS, Annecy-le-Vieux, France}
38: \newcommand{\APC}{APC, Paris, France}
39: \newcommand{\DIAS}{Dublin Institute for Advanced Studies, Ireland}
40: \newcommand{\LSW}{Landessternwarte, Universit\"at Heidelberg, K\"onigstuhl, Germany}
41: \newcommand{\Montpellier}{Laboratoire de Physique Th\'eorique et Astroparticules, IN2P3/CNRS, Universit\'e Montpellier II, Montpellier, France}
42: \newcommand{\Saclay}{DAPNIA/DSM/CEA, CE Saclay, Gif-sur-Yvette, France}
43: \newcommand{\Grenoble}{
44: Laboratoire d'Astrophysique de Grenoble, INSU/CNRS, Universit\'e Joseph Fourier, Grenoble, France}
45: \newcommand{\Tuebingen}{Institut f\"ur Astronomie und Astrophysik, Universit\"at T\"ubingen, Germany}
46: \newcommand{\Jussieu}{LPNHE, IN2P3/CNRS, Universit\'es Paris VI \& VII, France}
47: \newcommand{\Prague}{Institute of Particle and Nuclear Physics, Charles University, Prague, Czech Republic}
48: \newcommand{\Bochum}{Institut f\"ur Theoretische Physik, Lehrstuhl IV: Weltraum und Astrophysik, Ruhr-Universit\"at Bochum, Germany
49: }\newcommand{\Namibia}{University of Namibia, Windhoek, Namibia}
50: \newcommand{\Erlangen}{Universit\"at Erlangen-N\"urnberg, Physikalisches Institut, Germany}
51: \newcommand{\LEA}{European Associated Laboratory for Gamma-Ray Astronomy, jointly supported by CNRS and MPG}
52: \newcommand{\PolandA}{Obserwatorium Astronomiczne, Uniwersytet Jagiello\'nski, Krak\'ow, Poland}
53: \newcommand{\PolandB}{Nicolaus Copernicus Astronomical Center, Warsaw, Poland}
54: \newcommand{\pks}{PKS\,2155$-$304}
55:
56: %\slugcomment{Not to appear in Nonlearned J.}
57:
58: \shorttitle{An Exceptional VHE Gamma-Ray Flare of PKS\,2155$-$304}
59: \shortauthors{Aharonian et al.}
60:
61: \begin{document}
62: \title{An Exceptional VHE Gamma-Ray Flare of PKS\,2155$-$304}
63:
64: %% Use \author, \affil, and the \and command to format
65: %% author and affiliation information.
66: %% Note that \email has replaced the old \authoremail command
67: %% from AASTeX v4.0. You can use \email to mark an email address
68: %% anywhere in the paper, not just in the front matter.
69: %% As in the title, use \\ to force line breaks.
70:
71: \author{F. Aharonian\altaffilmark{1,2,14},
72: A.G.~Akhperjanian\altaffilmark{3},
73: A.R.~Bazer-Bachi\altaffilmark{4},
74: B.~Behera\altaffilmark{15},
75: M.~Beilicke\altaffilmark{5},
76: W.~Benbow\altaffilmark{2},
77: D.~Berge\altaffilmark{2,a},
78: K.~Bernl\"ohr\altaffilmark{2,6},
79: C.~Boisson\altaffilmark{7},
80: O.~Bolz\altaffilmark{2},
81: V.~Borrel\altaffilmark{4},
82: T.~Boutelier\altaffilmark{18},
83: I.~Braun\altaffilmark{2},
84: E.~Brion\altaffilmark{8},
85: A.M.~Brown\altaffilmark{9},
86: R.~B\"uhler\altaffilmark{2},
87: I.~B\"usching\altaffilmark{10},
88: T.~Bulik\altaffilmark{25},
89: S.~Carrigan\altaffilmark{2},
90: P.M.~Chadwick\altaffilmark{9},
91: A.C.~Clapson\altaffilmark{2},
92: L.-M.~Chounet\altaffilmark{11},
93: G.~Coignet\altaffilmark{12},
94: R.~Cornils\altaffilmark{5},
95: L.~Costamante\altaffilmark{2,26},
96: B.~Degrange\altaffilmark{11},
97: H.J.~Dickinson\altaffilmark{9},
98: A.~Djannati-Ata\"i\altaffilmark{13},
99: W.~Domainko\altaffilmark{2},
100: L.O'C.~Drury\altaffilmark{14},
101: G.~Dubus\altaffilmark{11},
102: J.~Dyks\altaffilmark{25},
103: K.~Egberts\altaffilmark{2},
104: D.~Emmanoulopoulos\altaffilmark{15},
105: P.~Espigat\altaffilmark{13},
106: C.~Farnier\altaffilmark{16},
107: F.~Feinstein\altaffilmark{16},
108: A.~Fiasson\altaffilmark{16},
109: A.~F\"orster\altaffilmark{2},
110: G.~Fontaine\altaffilmark{11},
111: Seb.~Funk\altaffilmark{6},
112: S.~Funk\altaffilmark{2},
113: M.~F\"u{\ss}ling\altaffilmark{6},
114: Y.A.~Gallant\altaffilmark{16},
115: B.~Giebels\altaffilmark{11},
116: J.F.~Glicenstein\altaffilmark{8},
117: B.~Gl\"uck\altaffilmark{17}
118: P.~Goret\altaffilmark{8},
119: C.~Hadjichristidis\altaffilmark{9},
120: D.~Hauser\altaffilmark{2},
121: M.~Hauser\altaffilmark{15},
122: G.~Heinzelmann\altaffilmark{5},
123: G.~Henri\altaffilmark{18},
124: G.~Hermann\altaffilmark{2},
125: J.A.~Hinton\altaffilmark{2,15,b},
126: A.~Hoffmann\altaffilmark{19},
127: W.~Hofmann\altaffilmark{2},
128: M.~Holleran\altaffilmark{10},
129: S.~Hoppe\altaffilmark{2},
130: D.~Horns\altaffilmark{19},
131: A.~Jacholkowska\altaffilmark{16},
132: O.C.~de~Jager\altaffilmark{10},
133: E.~Kendziorra\altaffilmark{19},
134: M.~Kerschhaggl\altaffilmark{6},
135: B.~Kh\'elifi\altaffilmark{11,2},
136: Nu.~Komin\altaffilmark{16},
137: K.~Kosack\altaffilmark{2},
138: G.~Lamanna\altaffilmark{12},
139: I.J.~Latham\altaffilmark{9},
140: R.~Le Gallou\altaffilmark{9},
141: A.~Lemi\`ere\altaffilmark{13},
142: M.~Lemoine-Goumard\altaffilmark{11},
143: J.-P.~Lenain\altaffilmark{7},
144: T.~Lohse\altaffilmark{6},
145: J.M.~Martin\altaffilmark{7},
146: O.~Martineau-Huynh\altaffilmark{20},
147: A.~Marcowith\altaffilmark{16},
148: C.~Masterson\altaffilmark{14},
149: G.~Maurin\altaffilmark{13},
150: T.J.L.~McComb\altaffilmark{9},
151: R.~Moderski\altaffilmark{25},
152: E.~Moulin\altaffilmark{16,8},
153: M.~de~Naurois\altaffilmark{20},
154: D.~Nedbal\altaffilmark{21},
155: S.J.~Nolan\altaffilmark{9},
156: J-P.~Olive\altaffilmark{4},
157: K.J.~Orford\altaffilmark{9},
158: J.L.~Osborne\altaffilmark{9},
159: M.~Ostrowski\altaffilmark{24},
160: M.~Panter\altaffilmark{2},
161: G.~Pedaletti\altaffilmark{15},
162: G.~Pelletier\altaffilmark{18},
163: P.-O.~Petrucci\altaffilmark{18},
164: S.~Pita\altaffilmark{13},
165: G.~P\"uhlhofer\altaffilmark{15},
166: M.~Punch\altaffilmark{13},
167: S.~Ranchon\altaffilmark{12},
168: B.C.~Raubenheimer\altaffilmark{10},
169: M.~Raue\altaffilmark{5},
170: S.M.~Rayner\altaffilmark{9},
171: M.~Renaud\altaffilmark{2},
172: J.~Ripken\altaffilmark{5},
173: L.~Rob\altaffilmark{21},
174: L.~Rolland\altaffilmark{8},
175: S.~Rosier-Lees\altaffilmark{12},
176: G.~Rowell\altaffilmark{2,c},
177: B.~Rudak\altaffilmark{25},
178: J.~Ruppel\altaffilmark{22},
179: V.~Sahakian\altaffilmark{3},
180: A.~Santangelo\altaffilmark{19},
181: L.~Saug\'e\altaffilmark{18},
182: S.~Schlenker\altaffilmark{6},
183: R.~Schlickeiser\altaffilmark{22},
184: R.~Schr\"oder\altaffilmark{22},
185: U.~Schwanke\altaffilmark{6},
186: S.~Schwarzburg \altaffilmark{19},
187: S.~Schwemmer\altaffilmark{15},
188: A.~Shalchi\altaffilmark{22},
189: H.~Sol\altaffilmark{7},
190: D.~Spangler\altaffilmark{9},
191: {\L}.~Stawarz\altaffilmark{24}
192: R.~Steenkamp\altaffilmark{23},
193: C.~Stegmann\altaffilmark{17},
194: G.~Superina\altaffilmark{11},
195: P.H.~Tam\altaffilmark{15},
196: J.-P.~Tavernet\altaffilmark{20},
197: R.~Terrier\altaffilmark{13},
198: C.~van~Eldik\altaffilmark{2},
199: G.~Vasileiadis\altaffilmark{16},
200: C.~Venter\altaffilmark{10},
201: J.P.~Vialle\altaffilmark{12},
202: P.~Vincent\altaffilmark{20},
203: M.~Vivier\altaffilmark{8},
204: H.J.~V\"olk\altaffilmark{2},
205: F.~Volpe\altaffilmark{11},
206: S.J.~Wagner\altaffilmark{15},
207: M.~Ward\altaffilmark{9},
208: A.~Zdziarski\altaffilmark{25}}
209:
210: \altaffiltext{1}{Correspondence and request for material should be
211: addressed to W. Benbow, \myemail, B. Giebels, \hisemail}
212: \altaffiltext{2}{\MPIK}
213: \altaffiltext{3}{\Yerevan}
214: \altaffiltext{4}{\Tolouse}
215: \altaffiltext{5}{\Hamburg}
216: \altaffiltext{6}{\Berlin}
217: \altaffiltext{7}{\Meudon}
218: \altaffiltext{8}{\Saclay}
219: \altaffiltext{9}{\Durham}
220: \altaffiltext{10}{\Potch}
221: \altaffiltext{11}{\LLR}
222: \altaffiltext{12}{\Annecy}
223: \altaffiltext{13}{\APC}
224: \altaffiltext{14}{\DIAS}
225: \altaffiltext{15}{\LSW}
226: \altaffiltext{16}{\Montpellier}
227: \altaffiltext{17}{\Erlangen}
228: \altaffiltext{18}{\Grenoble}
229: \altaffiltext{19}{\Tuebingen}
230: \altaffiltext{20}{\Jussieu}
231: \altaffiltext{21}{\Prague}
232: \altaffiltext{22}{\Bochum}
233: \altaffiltext{23}{\Namibia}
234: \altaffiltext{24}{\PolandA}
235: \altaffiltext{25}{\PolandB}
236: \altaffiltext{26}{\LEA}
237: \altaffiltext{a}{now at CERN, Geneva, Switzerland}
238: \altaffiltext{b}{now at University of Leeds, UK}
239: \altaffiltext{c}{now at University of Adelaide, Australia}
240:
241:
242: \begin{abstract}
243: The high-frequency peaked BL\,Lac PKS\,2155$-$304 at redshift
244: $z=0.116$ is a well-known VHE ($>$100 GeV) $\gamma$-ray emitter.
245: Since 2002 its VHE flux has been monitored using the
246: H.E.S.S. stereoscopic array of imaging atmospheric-Cherenkov
247: telescopes in Namibia. During the July 2006 dark period, the average
248: VHE flux was measured to be more than ten times typical values
249: observed from the object. This article focuses solely on an extreme
250: $\gamma$-ray outburst detected in the early hours of July 28, 2006
251: (MJD 53944). The average flux observed during this outburst is
252: I($>$200 GeV) = (1.72$\pm$$0.05_{\rm stat}$$\pm$$0.34_{\rm syst}$)
253: $\times$ 10$^{-9}$ cm$^{-2}$ s$^{-1}$, corresponding to $\sim$7 times
254: the flux, I($>$200 GeV), observed from the Crab Nebula. Peak fluxes
255: are measured with one-minute time scale resolution at more than twice
256: this average value. Variability is seen up to $\sim$600 s in the
257: Fourier power spectrum, and well-resolved bursts varying on time
258: scales of $\sim$200 seconds are observed. There are no strong
259: indications for spectral variability within the data. Assuming the
260: emission region has a size comparable to the Schwarzschild radius of a $\sim$$
261: 10^9\,M_\odot$ black hole, Doppler factors greater than 100 are
262: required to accommodate the observed variability time scales.
263:
264: \end{abstract}
265:
266: \keywords{Galaxies: active --
267: BL Lacertae objects: Individual: PKS\,2155$-$304 --
268: Gamma rays: observations}
269: \section{Introduction}
270:
271: Flux variability studies provide a strong probe into the physical
272: processes of the innermost regions of Active Galactic Nuclei (AGN).
273: Although the broad-band emission from all AGN is highly variable, the
274: most extreme flux variability, i.e. largest magnitude and shortest
275: time scale, is observed from a class of AGN known as blazars. As a
276: result blazar variability studies are crucial to unraveling the
277: mysteries of AGN. Over a dozen blazars have been detected so far at
278: VHE energies. In the Southern Hemisphere, PKS\,2155$-$304 is
279: generally the brightest blazar at these energies, and is probably the
280: best-studied at all wavelengths. The VHE flux observed
281: \cite{HESS_2155A} from PKS\,2155$-$304 is typically of the order
282: $\sim$15\% of the Crab Nebula flux above 200 GeV. The highest flux
283: previously measured in one night is approximately four times this
284: value and clear VHE-flux variability has been observed on daily time
285: scales. The most rapid flux variability measured for this source is
286: 25\,min~\cite{HESS_2155B}, occurring at X-ray energies. The fastest
287: variation published from any blazar, at any wavelength, is an event lasting
288: $\sim$800\,s where the X-ray flux from Mkn\,501 varied by 30\%
289: \cite[Xue \& Cui][]{MKN501_dispute}\footnote{Xue \& Cui~\cite{MKN501_dispute}
290: also demonstrate that a 60\% X-ray flux increase in $\sim$200\,s observed
291: \cite{MKN501_flare} from Mkn\,501 is likely an artifact.},
292: while at VHE energies doubling timescales as fast as
293: $\sim$15 minutes have been observed from Mkn 421 \cite{Gaidos_Mkn421}.
294:
295: The High Energy Stereoscopic System \cite[H.E.S.S.;][]{HESS1}
296: is used to study VHE $\gamma$-ray emission
297: from wide variety of astrophysical objects.
298: As part of the normal H.E.S.S. observation program
299: the flux from known VHE AGN is monitored regularly
300: to search for bright flares. During such flares,
301: the unprecedented sensitivity of H.E.S.S.
302: (5 standard deviation, $\sigma$,
303: detection in $\sim$30\,s for a Crab Nebula flux source at
304: 20$^{\circ}$ zenith angle) enables studies of VHE-flux variability on
305: time scales of a few tens of seconds. During the July 2006 dark period,
306: the average VHE flux observed by H.E.S.S. from PKS\,2155$-$304 was
307: more than ten times its typical value. In particular,
308: an extremely bright flare of PKS\,2155$-$304 was
309: observed in the early hours of July 28, 2006 (MJD 53944).
310: This article focuses solely on this particular flare.
311: The results from other H.E.S.S. observations of
312: PKS\,2155$-$304 from 2004 through 2006 will be published elsewhere.
313:
314: \section{Results from MJD 53944}
315: \label{sect:results}
316:
317: A total of three observation runs ($\sim$28 min each) were
318: taken on PKS\,2155$-$304 in the early hours\footnote{The three runs began at
319: 00:35, 01:06 and 01:36 UTC, respectively.} of MJD 53944. These data
320: entirely pass the standard H.E.S.S. data-quality selection criteria,
321: yielding an exposure of 1.32\,h live time at a mean zenith angle of
322: 13$^{\circ}$. The standard H.E.S.S. calibration \cite{calib_paper}
323: and analysis tools \cite{std_analysis} are used to extract the results
324: shown here. As the observed signal is exceptionally strong, the
325: event-selection criteria \cite{std_analysis} are performed using the
326: {\it loose cuts}, instead of the {\it standard cuts}, yielding an
327: average post-analysis energy threshold of 170 GeV. The {\it loose cuts}
328: are selected since they have a lower energy threshold and
329: higher $\gamma$-ray and background acceptance. The
330: higher acceptances avoid low-statistics issues
331: with estimating the background and significance
332: on short time scales, thus simplifying the analysis.
333: The on-source data are taken from a circular region of radius
334: $\theta_{cut}=0.2^{\circ}$ centered on PKS\,2155$-$304, and the
335: background (off-source data) is estimated using
336: the {\it Reflected-Region} method \cite{bgmodel_paper}.
337:
338: A total of 12480 on-source events and 3296 off-source events
339: are measured with an on-off normalization of 0.215.
340: The observed excess is 11771 events ($\sim$2.5Hz),
341: corresponding to a significance of 168$\sigma$ calculated following the
342: method of Equation (17) in Li \& Ma~\cite{lima}. It should be noted that
343: use of the {\it standard cuts} also yields a strong excess
344: (6040 events, 159$\sigma$) and results (i.e. flux, spectrum,
345: variability) consistent with those detailed later.
346:
347: \subsection{Flux Variability}
348: The average integral flux above 200 GeV observed from PKS\,2155$-$304
349: is I($>$200 GeV) =
350: (1.72$\pm$$0.05_{\rm stat}$$\pm$$0.34_{\rm syst}$)$\,\times\,$10$^{-9}$\,cm$^{-2}$\,s$^{-1}$,
351: equivalent to $\sim$7 times the I($>$200 GeV) observed
352: from the Crab Nebula
353: \cite[I$_{\mathrm{Crab}}$;][]{hess_crab}. Figure~\ref{flux_lc_1min}
354: shows I($>$200 GeV), binned in one-minute intervals, versus time. The
355: fluxes in this light curve range from 0.65 I$_{\mathrm{Crab}}$ to 15.1
356: I$_{\mathrm{Crab}}$, and their fractional root mean square (rms)
357: variability amplitude \cite{rms_noise_ref}
358: is F$_{\rm var}=0.58\pm0.03$. This is $\sim$2
359: times higher than archival X-ray
360: variability \cite[Zhang et al.][]{zhang1999,zhang2005}.
361: The Fourier power spectrum
362: calculated from Figure~\ref{flux_lc_1min}
363: is shown in Figure~\ref{fourier_power}. The error on the power
364: spectrum is the 90\% confidence interval estimated
365: from $10^4$ simulated light curves. These curves
366: are generated by adding a random constant to
367: each individual flux point, where this constant is
368: taken randomly from a Gaussian distribution
369: with a dispersion equal to the error of the respective point.
370: The average power expected when the measurement error
371: dominates is shown as a dashed line
372: \cite[see the Appendix in][]{rms_noise_ref}. There is power
373: significantly above the measurement noise level up to $1.6 \times
374: 10^{-3}\,{\rm Hz}$ ($600\,{\rm s}$). The power spectrum also shows that most of the power is at low
375: frequencies. The grey shaded area shows the 90\% confidence level
376: obtained by simulating $10^4$ light curves with a power-law Fourier
377: spectrum $P_\nu\propto \nu^{-2}$ \cite{timmer} and a random Gaussian
378: error as above. The power spectrum derived from the data is thus
379: compatible with a light curve generated by a stochastic process with a
380: power-law Fourier spectrum of index -2. An index of -1 produces too
381: much power at high frequencies and is rejected. These power spectra are
382: remarkably similar to those derived in
383: X-rays \cite[Zhang et al.][]{zhang1999} from the same source.
384:
385: Rapid variability is clearly visible in substructures that appear in
386: the light curve, with even shorter rise and decay time scales than found
387: in the Fourier analysis. In
388: order to quantify those time scales, the light curve is considered as
389: consisting of a series of bursts, which is common for AGN and
390: $\gamma$-ray bursts (GRBs). The ``generalized Gaussian'' shape from
391: Norris et al.~\cite{norris} is used to characterize these bursts,
392: where the burst intensity is described by: ${\rm I}(t) = A \exp [
393: -(|t-t_{\rm max}|/\sigma_{\rm r,d})^\kappa]$, where $t_{\rm max}$ is
394: the time of the burst's maximum intensity (A); $\sigma_{\rm r}$ and
395: $\sigma_{\rm d}$ are the rise ($t<t_{\rm max}$) and decay ($t>t_{\rm
396: max}$) time constants, respectively; and $\kappa$ is a measure of the
397: burst's sharpness. The rise and decay times, from half to maximum
398: amplitude, are $\tau_{r,d}=[\ln 2]^{1/\kappa}\sigma_{r,d}$. A peak
399: finding tool, using a Markov chain algorithm \cite{mor02}, selected
400: five significant bursts.
401: A function consisting of a superposition of an identical
402: number of bursts plus a constant signal
403: was fit\footnote{The Markov chain burst positions
404: were used to initialize $t_{\rm max}$ for each burst. All parameters
405: are left free in the fit.} to the data.
406: The best fit has a $\chi^2$ probability of
407: 20\% and the fit parameters are shown in
408: Table~\ref{burst_info}. Interestingly, there is a marginal trend for
409: $\kappa$ to increase with subsequent bursts,
410: making them less sharp, as the flare
411: progresses. The $\kappa$ values
412: are close to the bulk of those found by Norris et al.~\cite{norris},
413: but the time scales measured here are two orders of
414: magnitude larger.
415:
416: \begin{table}
417: \caption{The results of the best $\chi^2$ fit of the superposition
418: of five bursts and a constant to the
419: data shown in Figure~\ref{flux_lc_1min}. The constant term is
420: $0.27\pm0.03 \times 10^{-9}\,{\rm cm}^{-2}\,{\rm s}^{-1}$ (1.1
421: ${\rm I}_{\rm Crab}$).\label{burst_info}}
422: \centering
423: \begin{tabular}{ccccc}
424: \tableline\tableline
425: $t_{\rm max}$ & $A$ &$\tau_{\rm r}$ & $\tau_{\rm d}$ & $\kappa$ \\
426: $[$min$]$ & $[10^{-9}\,{\rm cm}^{-2}\,{\rm
427: s}^{-1}]$ & [s] & [s] & \\
428: \tableline\\
429: 41.0 & 2.7$\pm$0.2 & 173$\pm$28 & 610$\pm$129 & 1.07$\pm$0.20\\
430: 58.8 & 2.1$\pm$0.9 & 116$\pm$53 & 178$\pm$146 & 1.43$\pm$0.83\\
431: 71.3 & 3.1$\pm$0.3 & 404$\pm$219 & 269$\pm$158 & 1.59$\pm$0.42\\
432: 79.5 & 2.0$\pm$0.8 & 178$\pm$55 & 657$\pm$268 & 2.01$\pm$0.87\\
433: 88.3 & 1.5$\pm$0.5 & 67$\pm$44 & 620$\pm$75 & 2.44$\pm$0.41\\
434: \tableline\\
435: \end{tabular}
436: \end{table}
437:
438: During both the first two bursts there is clear doubling of the flux
439: within $\tau_{r}$. Such doubling is sometimes used as a
440: characteristic time scale of flux variability. For compatibility with
441: such estimators, the definition of doubling time,
442: $T_2 = |{\rm I}_{ij} \Delta T / \Delta {\rm I}|$, from Zhang
443: et al.~\cite{zhang1999} was used\footnote{Only values of $T_2$ with
444: less than 30\% uncertainty are considered.}. Here, $\Delta T = T_j - T_i$,
445: $\Delta {\rm I} = {\rm I}_j - {\rm I}_i$, $
446: {\rm {\rm I}}_{ij} = ({\rm I}_j + {\rm I}_i)/2$, with $T$
447: and I being the time and flux, respectively, of any pair
448: of points in the light curve.
449: The fastest $T_2=224\pm60\,{\rm s}$ is
450: compatible with the fastest significant time scale found by the
451: Fourier transform. Averaging the five lowest $T_2$
452: values yields $330\pm40\,{\rm s}$.
453:
454: The variability time scales of these bursts are among
455: \cite[see also][]{MAGIC_501} the fastest ever
456: seen in a blazar, at any wavelength, and are almost an order of
457: magnitude smaller than previously observed from this object. It
458: should be noted that similar time scales are found with even smaller
459: binning (e.g. 20\,s) of the H.E.S.S. light curve, and that many checks
460: of the data quality were undertaken to ensure that the flux variations
461: cannot be the result of background fluctuations, atmospheric events,
462: etc. In addition, all the results have been verified using an
463: independent calibration method and alternative analysis
464: techniques.
465:
466: \subsection{Spectral Analysis}
467:
468: Figure~\ref{avg_spectrum} shows the time-averaged photon
469: spectrum for these data. The data are well fit, $\chi^2=17.1$ for 13
470: degrees of freedom (d.o.f.), by a broken power-law function:
471:
472: \noindent
473: $E < E_{\mathrm B}: \frac{\rm dN}{\rm dE} = I_{\circ} \hspace{0.5ex} \left(\frac{E}{\rm 1\,TeV}\right)^{-\Gamma_{1}} \hspace{0.5ex}$
474:
475: \noindent
476: $E > E_{\mathrm B}: \frac{\rm dN}{\rm dE} = I_{\circ} \hspace{0.5 ex} \left(\frac{E_{\mathrm B}}{\rm 1\,TeV}\right)^{(\Gamma_{2} - \Gamma_{1})} \left(\frac{E}{\rm 1\,TeV}\right)^{-\Gamma_{2}}$,
477:
478: \noindent
479: where $I_{\circ}\,$=\,(2.06$\pm$0.16$\pm$0.41)
480: $\times$ 10$^{-10}$ cm$^{-2}$\,s$^{-1}$\,TeV$^{-1}$,
481: $E_{\mathrm{B}}$\,=\,430$\pm$22$\pm$80 GeV,
482: $\Gamma_{1}$\,=\,2.71$\pm$0.06$\pm$0.10,
483: and $\Gamma_{2}$\,=\,3.53$\pm$0.05$\pm$0.10.
484: For each parameter, the two uncertainties are the
485: statistical and systematic values, respectively.
486: Fits to the data of either a simple power law
487: ($\Gamma$\,=\,3.19$\pm$0.02$\pm$0.10,
488: $\chi^2$\,=\,138, 15 d.o.f) or a power law with an exponential cut-off
489: ($\chi^2$\,=\,45, 14 d.o.f.) are not acceptable.
490: The time-averaged spectrum ($\Gamma$\,=\,3.32)
491: of \pks\ measured in 2003 \cite{HESS_2155A},
492: multiplied by the ratio (48.7) of I($>$200~GeV) from the
493: respective data sets, is also shown in Figure~\ref{avg_spectrum}.
494: Despite a factor of $\sim$50 change in flux there
495: is qualitatively little difference between the two spectra.
496: Indeed, fitting a broken power law to the
497: current data set, keeping $\Gamma_{1}$ and $\Gamma_{2}$
498: fixed to the values measured in 2003, yields a
499: value for $E_{\mathrm{B}}$ consistent with that
500: measured in 2003. The small difference is surprising since a
501: change of the spectral shape with varying flux levels, typically
502: hardening with increased flux, has often
503: been observed from blazars at X-ray
504: energies \cite[see, e.g.,][]{Xray_hardening},
505: as well as in the VHE domain \cite[see, e.g.,][]{VHE_hardening}.
506:
507: The high flux observed from PKS\,2155$-$304 allows the determination of
508: accurate photon spectra on time scales of the order of minutes.
509: Therefore, a simple search for temporal changes of
510: the VHE spectral shape within these data was performed. Spectra were determined
511: for consecutive data slices of 28 minutes (1 run), 10 minutes,
512: and 5 minutes. Fitting the time-average spectral shape, allowing only the
513: normalization ($I_{\circ}$) to vary, to these short-time-scale spectra
514: yields reasonable $\chi^2$ probabilities.
515: Thus, there are no strong indications
516: of fast spectral variability. However,
517: weak variations ($\Delta \Gamma$\,$<$\,0.2) are not ruled out.
518: A more sophisticated study of any fast spectral variations
519: within these data is beyond the scope of this
520: letter and will be published elsewhere.
521:
522: \section{Discussion}
523:
524: It is very likely that the electromagnetic emission in blazars
525: is generated in jets that are beamed
526: and Doppler-boosted toward the observer.
527: Superluminal expansions observed with VLBI \cite{piner}
528: provide evidence for moderate Doppler boosting in PKS\,2155$-$304.
529: Causality implies that $\gamma$-ray variability on a
530: time scale $t_{\rm var}$, with a Doppler factor\footnote{With
531: $\delta$ defined in the standard
532: way as $[\Gamma(1-\beta\cos\theta)]^{-1}$, where $\Gamma$ is the bulk
533: Lorentz factor of the plasma in the jet, $\beta = v/c$, and $\theta$
534: is the angle to the line of sight.} ($\delta$), is related to the
535: radius ($R$) of the emission zone by $R \leq ct_{\rm
536: var}\delta/(1+z)$. Conservatively using the best-determined rise time
537: (i.e. $\tau_r$ with the smallest error)
538: from Table~\ref{burst_info} for $t_{\rm var} = 173\pm28\,{\rm s}$
539: (note that this is similar to the
540: fastest $T_2$) limits the size of the emission region
541: to $R\delta^{-1} \leq 4.65 \times 10^{12}$ cm $\leq 0.31$ AU.
542:
543: The jets of blazars are believed to be powered by accretion onto a
544: supermassive black hole (SMBH). Thus accretion/ejection
545: properties are usually presumed to scale with the Schwarzschild radius
546: $R_{\rm S}$ of the SMBH, where $R_{\rm S} = 2GM/c^2$, which
547: is the smallest, most-natural size of the system
548: \cite[see, e.g.,][]{Blandford}.
549: Expressing the size $R$ of the $\gamma$-ray emitting region
550: in terms of $R_{\rm S}$, the variability time
551: scale limits its mass by $M \leq (c^3 t_{\rm var}\delta/2G(1+z)) R_{\rm S}/R
552: \sim 1.6\times10^7 M_\odot \delta R_{\rm S} / R$.
553: The reported\footnote{See Wurtz
554: et al.~\cite{wurtz} and $M_R>-23.1$ (for $h=0.5$) showing the need for
555: confirmation of this value.} host galaxy luminosity $M_R=-24.4$
556: \cite[Table 3 in][]{kotilainen} would imply a SMBH mass of order
557: 1$-$2$\times 10^9M_\odot$ ~\cite{bettoni2003}, and therefore,
558: $\delta\geq 60-120\,R/R_{\rm S}$. Emission regions of only a
559: few $R_{\rm S}$ would
560: require values of $\delta$ much greater than those typically
561: derived for blazars ($\delta$$\sim$10) and come close
562: to those used for GRBs, which would be a challenge to understand. For
563: example, the sub-parsec VHE $\gamma$-ray emitting plasma would have to
564: decelerate with a high efficiency to accommodate relatively small
565: Lorentz factors observed at parsec scales \cite{piner}. It is however
566: possible that the SMBH mass is over-estimated, reducing the $\delta$
567: constraint by the same factor, or that the variability has an
568: origin \cite[e.g., a geometric effect from jet bending
569: as discussed in][]{wagner1993} unrelated to the black hole.
570: Detailed modeling of the spectral energy distribution
571: of PKS\,2155$-$304, during the multiple VHE
572: flares observed by H.E.S.S. in the July 2006 dark period,
573: including simultaneous multi-frequency data, will
574: appear elsewhere.
575:
576: The VHE variability observed in this particular flaring episode
577: is the fastest ever observed from a blazar. While the
578: variability is a factor of five times faster than
579: previously measured from Mkn 421 \cite{Gaidos_Mkn421}, in terms of the
580: light-crossing time of the Schwarzschild
581: radius, $R_{\rm S}/c$, the variability of PKS\,2155$-$304 is
582: another factor of $\approx 6-12$ more constraining assuming a
583: $10^{8.22} M_\odot$ for Mkn 421 \cite{BH_Mass}.
584: It should also be noted that the
585: choice of a $\sim$3 minute variability time scale here is
586: conservative and that the light curve is strongly oversampled,
587: allowing for the first time in the VHE regime a detailed
588: statistical analysis of a flare, which shows remarkable similarity to
589: other longer duration events at X-ray energies.
590: From such rapid variability one must conclude that
591: either very large Doppler factors can
592: be present in AGN jets, or that the observed variability is not
593: connected to the central black hole, clearly showing the power
594: of Cherenkov-telescope arrays in probing the
595: internal mechanisms in BL\,Lacs.
596:
597: \acknowledgments
598:
599: The support of the Namibian authorities and of the University of Namibia
600: in facilitating the construction and operation of H.E.S.S. is gratefully
601: acknowledged, as is the support by the German Ministry for Education and
602: Research (BMBF), the Max Planck Society, the French Ministry for Research,
603: the CNRS-IN2P3 and the Astroparticle Interdisciplinary Programme of the
604: CNRS, the U.K. Particle Physics and Astronomy Research Council (PPARC),
605: the IPNP of the Charles University, the Polish Ministry of Science and
606: Higher Education, the South African Department of
607: Science and Technology and National Research Foundation, and by the
608: University of Namibia. We appreciate the excellent work of the technical
609: support staff in Berlin, Durham, Hamburg, Heidelberg, Palaiseau, Paris,
610: Saclay, and in Namibia in the construction and operation of the
611: equipment.
612:
613: \begin{thebibliography}{}
614: \bibitem[Aharonian et al. 2002]{VHE_hardening}
615: Aharonian, F., et al. (HEGRA Collaboration) 2002, A\&A, 393, 89
616: \bibitem[Aharonian et al. 2004]{calib_paper}
617: Aharonian, F., et al. (H.E.S.S. Collaboration) 2004,
618: Astropart. Phys., 22, 109
619: \bibitem[Aharonian et al. 2005a]{HESS_2155A} Aharonian, F., et al. (H.E.S.S. Collaboration) 2005a, A\&A, 430, 865
620: \bibitem[Aharonian et al. 2005b]{HESS_2155B} Aharonian, F., et al. (H.E.S.S. Collaboration) 2005b, A\&A, 442, 895
621: \bibitem[Aharonian et al. 2006]{hess_crab} Aharonian, F., et al. (H.E.S.S. Collaboration) 2006, A\&A, 457, 899
622: \bibitem[Albert et al. 2007]{MAGIC_501} Albert, J., et al. 2007, ApJ, in press [arXiv:astro-ph/0702008]
623: \bibitem[Benbow 2005]{std_analysis} Benbow, W. 2005, Proc. of Towards a Network of Atmospheric Cherenkov Detectors VII (Palaiseau), 163
624: \bibitem[Berge et al. 2007]{bgmodel_paper} Berge, D., Funk, S. \&
625: Hinton, J. 2007, A\&A, 466, 1219
626: \bibitem[Bettoni et al. 2003]{bettoni2003} Bettoni, D., Falomo, R.,
627: Fasano, G. \& Govoni, F. 2003, \aap, 399, 869
628: \bibitem[Blandford et al. 1982]{Blandford}
629: Blandford, R.D. \& Payne, D.G. 1982, MNRAS, 199, 883
630: \bibitem[Catanese \& Sambruna 2000]{MKN501_flare} Catanese, M. \&
631: Sambruna, R.M. 2000, \apj, 534, L39
632: \bibitem[Gaidos et al. 1996]{Gaidos_Mkn421}
633: Gaidos, J.A., et al. 1996, Nature, 383, 319
634: \bibitem[Giommi et al. 1990]{Xray_hardening}
635: Giommi, P., et al. 1990, \apj, 356, 432
636: \bibitem[Hinton 2004]{HESS1} Hinton, J. 2004, New Astron Rev, 48, 331
637: \bibitem[Kotilainen et al. 1998]{kotilainen} Kotilainen,
638: J.K., Falomo, R., \& Scarpa, R. 1998, A\&A, 336, 479
639: \bibitem[1983]{lima} Li, T. \& Ma, Y. 1983, \apj, 272, 317
640: \bibitem[Morhac 2000]{mor02} Morhac, M., et al. 2000, Nucl. Instrum. and
641: Methods Phys. Res. A, 443, 108
642: \bibitem[1996]{norris} Norris, J.P., et al. 1996, ApJ, 459, 393
643: \bibitem[Piner \& Edwards 2004]{piner} Piner, B.G. \& Edwards,
644: P.G. 2004, \apj, 600, 115
645: \bibitem[Timmer \& Koenig 1995]{timmer} Timmer, J. \& Koenig, M. 1995, \aap, 300, 707
646: \bibitem[Vaughan et al. 2003]{rms_noise_ref}
647: Vaughan, S., Edelson, R., Warwick, R.S. \& Uttley, P. 2003,
648: MNRAS, 345, 1271
649: \bibitem[2005]{MKN501_dispute} Xue, Y. \& Cui, W. 2005, \apj, 622, 160
650: \bibitem[Wagner et al. 1993]{wagner1993} Wagner, S.J., et al. 1993, \aap,
651: 271, 344
652: \bibitem[Woo et al. 2005]{BH_Mass} Woo, J.-H., et al. 2005, \apj, 631, 762
653: \bibitem[1996]{wurtz} Wurtz, R., Stocke, J.T., \& Yee,
654: H.K.C. 1996, ApJS, 103, 109
655: \bibitem[1999]{zhang1999} Zhang, Y.H., et al. 1999, \apj, 527, 719
656: \bibitem[Zhang et al. 2005]{zhang2005} Zhang, Y.H., et al. 2005, \apj, 629, 686
657: \end{thebibliography}
658:
659: \clearpage
660:
661: \begin{figure}
662: \plotone{f1.eps}
663: \caption{The integral flux above 200 GeV observed
664: from PKS\,2155$-$304 on MJD 53944 versus time. The data are binned
665: in 1-minute intervals. The horizontal line represents
666: I($>$200 GeV) observed \cite{hess_crab}
667: from the Crab Nebula. The curve is the fit
668: to these data of the superposition of five bursts (see text) and a
669: constant flux.\label{flux_lc_1min}}
670: \end{figure}
671:
672: %\clearpage
673:
674: \begin{figure}
675: \plotone{f2.eps}
676: \caption{The Fourier power spectrum of the
677: light curve and associated measurement error. The grey shaded area
678: corresponds to the 90\% confidence interval for a light curve with a power-law
679: Fourier spectrum $P_{\nu}\propto \nu^{-2}$. The horizontal line is the
680: average noise level (see text). \label{fourier_power}}
681: \end{figure}
682:
683: \clearpage
684:
685: \begin{figure}
686: \plotone{f3.eps}
687: \caption{The time-averaged spectrum observed
688: from PKS\,2155$-$304 on MJD 53944.
689: The dashed line is
690: the best $\chi^2$ fit of a broken power law to
691: the data. The solid line represents the fit to
692: the time-averaged spectrum of PKS\,2155$-$304
693: from 2003 \cite{HESS_2155A}
694: scaled by 48.7. Neither spectrum is
695: corrected \cite[see, e.g.,][]{HESS_2155B}
696: for the absorption of VHE $\gamma$-rays
697: on the Extragalactic Background Light.
698: \label{avg_spectrum}}
699: \end{figure}
700:
701: \end{document}
702: