2 %\documentclass[12pt]{article}
4 %\usepackage{longtable}
7 % Section on the EMCal step manager stepping parameters.
9 \subsection{Step Manger and Hits Creation}
10 The majority of time and effort associated with a detector Monte Carlo
11 is involved in the transport of the particles, one at a time typically,
12 though the detector geometry. This is handled by a routine typically called
13 the ``step manager''. This routine, in general, does a lot of stuff with
14 considerable help from other sub-packages. It must determine what size
15 step to make based on the distance to the next volume, the curvature of the
16 track, the probability of some non-continuum process occurring (an
17 interaction), and deal with particles no longer being transported
18 (dropping below cuts); computing the effects of all continuum process
19 (energy loss, fluctuations, and multiple scattering); and outputting,
20 when relevant, any information to the ``user''. The majority of these
21 tasks are common to all detectors and are therefor done for us
22 with the help of the geometrical modeler and/or simulation framework.
24 the outputting of information an EMCal specific {\bf StepManager} routine
25 located in the \texttt{\bf AliEMCALv1}\footnote{There are more than one
26 version of {\bf AliEMCALv1} depending on differences in geometry and
27 some physics. All are derived from the EMCal class \texttt{\bf AliEMCALv0}
28 which is derived from \texttt{\bf AliEMCAL}, which is derived from
29 \texttt{AliDetector} which is derived from \texttt{AliModule}.} module,
30 or equivalent is used. Information outputted by this routine are
31 called ``hits'', in the AliRoot terminology, and are written to a
32 file called \texttt{EMCAL.Hits.root}. Often in production simulations
33 this file will be deleted afters the digits are produced.
35 The EMCal StepManager is called from the Alice implementation of the
36 ROOT virtual Monte Carlo step manager, specifically the routine
37 \texttt{\bf AliMC::Stepping}. It inquires, from the transport engine and its
38 geometrical modeler, what material the presently transporting particle
39 is in and deals with a couple of remaining particle transport issues
40 and then calls the detector specific step manager routine. This decoding
41 is done quickly through an array of material ID numbers indexed to
42 their corresponding detectors. Consequently, each sub-detector must
43 have its own unique material definitions obeying the ALICE material
44 numbering conventions. In this way, the addition or absence of a
45 sub-detector is dynamically handled via the initialization of the
46 material/detector array and the \texttt{TObject} array of sub-detectors
47 (all derived from the \texttt{\bf AliModule} class). This initialization
48 is done by the \texttt{Config.C} script.
50 \subsubsection{The EMCal Step Manager}
51 The first difficulty faced in the EMCal step manager is the extremely large
52 number of tracks produced and all of their individual steps. Recording each
53 step location, momentum, energy loss, and the like, for all of those
54 shower particles would overwhelm most IO systems and create too much
55 data to try to deal with further on in the simulation. Yet we need the
56 particle transport engine to generate and transport the majority of
57 these shower particles, otherwise, the signals in the neighboring
58 towers would be grossly incorrect and any part of a shower which goes
59 beyond the EMCal would also not be dealt with properly. In much thinner
60 detectors, like the ITS, the particles parameters at each step is
61 recorded directly into the hits.
63 For each particle entering the EMCal, what we want to record is the
64 energy lost by it and all of its shower daughters and in which tower
65 this energy loss occurred (and only for the ``sensitive'' materials/volumes
66 in the towers). In fact we really only want to associate this tower-wise
67 energy loss to the ``primary'' particle. To do this, for as long as the
68 ``primary'' track hasn't changed and the present track is still in the
69 same tower, the signals are added together (by adding their hits together.
70 This is done in \texttt{\bf AliEMCALv1::AddHit}). The determination of the
71 ``primary'' parent particle isn't so difficult, but one need to
72 deal with a number of special cases, and search back though the
73 parentage tree in some cases.
75 This leads to the $2^{nd}$ major task of the EMCal step manger routine. It
76 must determine which tower the transported particle is in. This is very
77 dependent on the details of the geometry and how it has been coded.
78 We know which volume the particle is in, but since there are many
79 copies (of the directory like geometry structure. figure
80 \ref{fig:GeometryDirectory}) of the tower volume,
81 we also need to find the necessary copy index numbers associated with
82 the specific volume. This is easy to get from the geometric modeler
83 (ROOT's TGeo package in our case), but it can be non-trivial to
84 convert these numbers into the tower, module, super-module index
85 wanted by the following simulation and reconstruction routines.
86 The present geometry, where a single tower sized scintillator
87 volume has the lead radiators embedded into it, simplifies this
88 tower determination because there is only one sensitive tower
89 sized volume and not a lot of individual sheets of scintillator
94 \includegraphics[width=0.8\textwidth]{figures/EMCalGeometryStructure.pdf}
96 \caption{\label{fig:GeometryDirectory}
97 Here is shown a typical hierarchical geometry structure. This
98 is similar to a directory structure except at each level one or
99 more copies, including translation and rotation operators, of
100 the daughters can be specified.}
103 For the EMCal we do something a bit special (but not untypical
104 for a calorimeter using organic scintillators). We correct for
105 the diminished light output
106 due to the ionization produced by the particles proceeding it.
107 This is done by rescaling the energy deposited using Birk's
108 law, equation \ref{equation:Birks}, as copied from GEANT3's
109 \texttt{G3BRIRK} routine
110 \cite{GEANT3:documentatoin}. This can be switched on or off
111 from the EMCal creation section of \texttt{Config.C} via
112 the \texttt{fBirkC0} variable in \texttt{AliEMCAL} class\footnote{
113 A function needs to be added to this class to allow for setting
114 this value and the Birk's law constants \texttt{fBirkC1}
115 and \texttt{fBirkC2}}. There has been
116 some debate about the proper way to deal with this in the
117 collaboration, mostly dealing with the limitations of any
118 Monte Carlo which transports particles one at a time, but
119 it has been agreed that including such a correction is
120 better than none at all.
124 \frac{\Delta E_{deposited}}{1+C_1 \delta + C_2 \delta^2}
125 \label{equation:Birks} \\
127 \frac{1}{\rho}\frac{dE}{dx}\; \left[\frac{MeV\: cm^{2}}{g}\right]\nonumber \\
128 C_{1} & = & \left\{ \begin{array}{ll}
129 0.013\: \left[\frac{g}{MeV\;cm^{2}}\right] & Z=1 \\
130 0.00743\; \left[\frac{g}{MeV\; cm^{2}}\right] &
131 Z>1 \end{array}\right.\nonumber \\
132 C_{2} & = & 9.6\times 10^{-6}\;
133 \left[\frac{g^{2}}{MeV^{2}\: cm^{4}}\right] \nonumber
136 The remaining tasks of the EMCal step manager is mostly book-keeping.
137 We only want to go to all of this effort if there is energy
138 being deposited in the sensitive scintillator volume, and not the
139 lead radiators or other structural materials. All of the relevant
140 information for the EMCal hit needs to be gathered. Lastly,
141 the \texttt{\bf AliEMCALHit} class needs to be created within the
142 \texttt{TClonesArray} of EMCal hits. This leads to some convoluted
143 looking code involving the TClonesArray fHits, the new
144 operator and the \texttt{AliEMCALHit} copy constructor (see
145 \texttt{\bf AliEMCAL::AddHit}).
147 The structure of these \texttt{EMCALHit} class (data structure) is
148 simple. It starts with the \texttt{AliHit} information which
149 consists of the tTrack number of the track which entered
150 the EMCal and its x,y,z global position (in cm). The EMCal specific
151 derivation includes the absolute tower ID where the hit signal is
152 from, the energy deposited by the showering particles
153 originating from this track in that tower, and the relative time
154 (with respect to the initial event) when this energy was deposited,
155 the particle ID of the particle entering the EMCal, the entrance
156 energy of the particle entering the EMCal, and the energy and
157 momentum of the primary particle entering the EMCal.
159 Just a note, although not included in the code, the addition of
160 the signals from the APD, primarily due to neutrons interacting
161 with the APD, needs to be added. CMS has found that including
162 this effect measurably improves the response of their simulations.
163 This will require an addition to the EMCal step manager, but hopefully
164 not the \texttt{EMCALHit} structure.
166 \subsubsection{Step Manager and Monte Carlo Setting}
167 In the EMCal geometry description there are also settings done,
168 on a medium by medium basis, which are used in the non-EMCal
169 specific step manager code. In \texttt{AliEMCAL} where ever
170 a medium is defined (either by a call to \texttt{AliMedium}
171 or equivalently to a call to \texttt{TGeoMedium}) a list of
172 parameters must be given which effects the size of a step.
173 These parameters are given in Table \ref{tab:MediumParameers}.
176 \begin{longtable}{p{0.12\textwidth}p{0.1\textwidth}p{0.78\textwidth}}
177 \multicolumn{3}{l}{Table \ref{tab:MediumParameers}} \\
179 Type & Variable & Description \\ \hline
181 \multicolumn{3}{l}{\emph{Table \ref{tab:MediumParameers} continued}}\\
183 Type & Variable & Description \\
187 \multicolumn{3}{r}{\emph{Table \ref{tab:MediumParameers} continued
191 \caption{Parameters and flags defined in the EMCal geometry via
192 a call to \texttt{ALIMedium} or \texttt{TGeoMedium}.
193 Because we use a version of \texttt{GEANT3}
194 which has its geometrical modeler replaced by \texttt{TGeo}
195 geometry, they are the same. This is also true for
196 both \texttt{GEANT4}\cite{GEANT4} and
197 \texttt{Fluka}\cite{Fluka} particle
198 transport Monte Carlos. See figure
199 \ref{fig:emcalStepManager_ParticleStep} for
200 a geometrical description of some of these parameters.
201 This information comes from the
202 GEANT3 documentation CONS200-1 \cite{GEANT3:documentatoin}.
203 \label{tab:MediumParameers}.
206 Int\_t & isvol & Sensitive volume flag.\newline
207 0 Not a Sensitive volume.\newline
208 1 Sensitive volume. \\
209 Int\_t & ifield & Magnetic field flag. \newline
210 0 No magnetic field.\newline
211 -1 User decision in \texttt{guswim}. Not supported
213 1 Tracking performed with Runge Kutta.\newline
214 2 Tracking performed with helix. \newline
215 3 constant magnetic field along z.\\
216 Float\_t & fieldm & Maximum magnetic field [kG].\\
217 Float\_t & tmaxfd & Maximum deflection angle due to magnetic
219 Float\_t & stemax & Maximum step allowed [cm].\\
220 Float\_t & deemax & Maximum fractional energy loss in one step.\newline
221 $dee=\frac{\Delta E}{E_{k}}$\\
222 Float\_t & epsil & Tracking precision [cm].\newline
223 This effects transition to new volumes.\\
224 Float\_t & stmin & Minimum step due to continuous processes [cm].\newline
225 This must be set to 0 so that \texttt{GEANT3} will
226 computing it correctly. Not doing so will adversely
227 effect the simulation.\\
229 \label{tab:MediumParameers}
233 This isn't the whole story in regards to the step manager. There
234 are a number of things which we need to set/control that don't
235 appear in any EMCal code, but are dealt with in the transport
236 engines part of the step manager. Such settings and controls
237 are very dependent on the specific transport Monte Carlo being
238 used. All of these settings and controls have ALICE wide
239 default values, but most of them we will need to change to
240 get optimal performance and accuracy from our EMCal simulation.
242 settings are settable for specific materials. If a material
243 does not have a set of switches or settings set, the ALICE
244 wide defaults are used.
246 \begin{tabular*}{\textwidth}[ht]{p{.5\textwidth}p{0.5\textwidth}}
247 \includegraphics[width=0.5\textwidth]{figures/EMCalMCStep.pdf}
248 \label{fig:emcalStepManager_ParticleStep}
251 Figure \ref{fig:emcalStepManager_ParticleStep}
252 An exaggerated Monte Carlo step showing some of the considerations associated
253 with transporting a Monte Carlo particle through a step. One step between
254 $s_{i}$ to $s_{i+1}$ is shown in the dotted (red) line. The dashed (blue)
255 line shows the step length taken due to a magnetic field. The solid (black)
256 line show what the particle path might really be. A new/different volume
257 is shown as the hashed (yellow) area. A new momentum and energy are
258 computed at the end of each step taking into account the energy loss
259 and multiple scattering. Also indicated is the deviation in the step
260 due to an applied magnetic field $t$, and the precision with which
261 the step has missed the other volume.
266 \paragraph{GEANT3 Switches and Settings}
269 GEANT3 was the first particle transport Monte Carlo integrated into
270 AliRoot and ROOT's virtual Monte Carlo and so has some of the
271 oldest and simplest interfaces. For simulation, AliRoot sets many
272 default settings and switches. This is done in \texttt{Config.C} which
273 you can find in \texttt{\$ALICE\_ROOT/macros}. There you will see a
274 number of line of the form \texttt{gMC->SetProcess(char *name,int value)}.
275 The switch names and there ALICE default values are shown in table
276 \ref{table:GEANT3PhysicsFlags}. There are limits both to the computer's
277 capabilities in dealing with the number of particles to transport and
278 with the physics models used by GEANT3. Consequently there are ``cuts''
279 used to stop the transport of particles which are below some energy
280 or are taking too long. The ALICE wide default values are also set
281 in \texttt{Config.C} using the function \texttt{gMC->SetCut(char *name,
282 double value)}. All of these values, and those included in the
283 \texttt{galice.cuts} file are listed in table \ref{table:GEANT3PhysicsLimits}.
285 The \texttt{galice.cuts} file has a fixed format as indicated in
286 \texttt{\bf AliMC::ReadTransPar} function. In this file, lines
287 staring with an ``*'' are ignored. The remaining lines are required to
288 contain, in order separated by one or more spaces, Detector\_Name,
289 Detector's\_media\_number, and then the numbered cuts and flags listed in
290 tables \ref{table:GEANT3PhysicsFlags} and \ref{table:GEANT3PhysicsLimits}.
292 \begin{longtable}{p{0.12\textwidth}p{0.1\textwidth}p{0.78\textwidth}}
293 \multicolumn{3}{l}{Table \ref{table:GEANT3Switchs}} \\
295 Switch & \small ALICE Default values & Description \\ \hline
297 \multicolumn{3}{l}{\emph{Table \ref{table:GEANT3Switchs} continued}}\\
299 Switch & \small ALICE Default value & Description \\
303 \multicolumn{3}{r}{\emph{Table \ref{table:GEANT3Switchs} continued
308 %\multicolumn{3}{p{0.95\textwidth}}{Table \ref{table:GEANT3Switches}:
309 \label{table:GEANT3Switchs}GEANT3 physics process flags.
310 These flags can be set on a
311 material by material basis. The ALICE Default values are
312 set in the \texttt{Config.C} file uses the
313 \texttt{gMC->SetProcess}
314 function. The setting of these specific flags for any
315 specific material is done in
316 \texttt{\$ALICE\_ROOT/data/galice.cuts}
317 file. The number on the left of the switch name is the
318 column in the \texttt{galice.cuts} file that this switch
319 is expected to be found. This information comes from the
320 GEANT3 documentation PHYS001-3 \cite{GEANT3:documentatoin}.
324 13 ANNI & 1 & Positron annihilation. The $e^+$ is stopped.\newline
325 0 No position annihilation.\newline
326 1 Positron annihilation with generation of $\gamma$.\newline
327 2 Positron annihilation without generation of $\gamma$.\\
329 14 BREM & 1 & bremsstrahlung. The interaction particle ($e^-$, $e^+$,
330 $\mu^-$, $\mu^+$) is stopped.\newline
331 0 No bremsstrahlung. \newline
332 1 bremsstrahlung with generation of $\gamma$.\newline
333 2 bremsstrahlung without generation of $\gamma$.\\
335 15 COMP & 1 & Compton scattering.\newline
336 0 No Compton scattering.\newline
337 1 Compton scattering with generation of $e^-$. \newline
338 2 Compton scattering without generation of $e^-$.\\
340 16 DCAY & 1 & Decay in flight. The decaying particles stops. \newline
341 0 No decay in flight \newline
342 1 Decay in flight with generation of secondaries \newline
343 2 Decay in flight without generation of secondaries \\
345 17 DRAY & 0 & $\delta$-ray production.\newline
346 0 No $\delta$-ray production.\newline
347 1 $\delta$-ray production with generation of $e^-$.\newline
348 2 $\delta$-ray production without generation of $e^-$.\\
350 18 HADR & 1 & Hadronic interactions. The particle is stopped in case
351 of inelastic interactions, while it is not stopped in case
352 of elastic interactions.\newline
353 0 No hadronic interactions.\newline
354 1 Hadronic interactions with generation of secondaries.\newline
355 2 Hadronic interactions without generation of
357 $>2$ can be used in the user code \texttt{GUPHAD} and
358 \texttt{GUHADR} to choose
359 a hadronic package. These values have no effect on the
360 hadronic packages themselves. Not supported in AliRoot.\\
362 19 LOSS & 2 & Continuous energy loss.\newline
363 0 No continuous energy loss, DRAY is forced to 0.\newline
364 1 Continuous energy loss with generation of $\delta$-rays
365 which have an energy above DCUTE and restricted
366 Landau-fluctuations\cite{LandauFluct} for $\delta$-rays
368 energy below DCUTE (no $\delta$-ray produced).\newline
369 2 Continuous energy loss without generation of $\delta$-rays
370 and full Landau-Vavilov-Gauss\cite{LandauVavilov}
371 fluctuations. In this case
372 DRAY is forced to 0 to avoid double counting of
373 fluctuations.\newline
374 3 Same as 1, kept for backwards compatibility.\newline
375 4 Energy loss without fluctuations. The value obtained
376 from the tables is used directly.\\
378 20 MULS & 1 & Multiple scattering.\newline
379 0 No multiple scattering.\newline
380 1 Multiple scattering according to Moliere\cite{Moiere}
382 2 Same as 1. Kept for backwards compatibility.\newline
383 3 Pure Gaussian scattering according to the Rossi
384 formula\cite{Rossi}.\\
386 21 PAIR & 1 & Pair production. The interacting $\gamma$ is
388 0 No pair production. \newline
389 1 Pair production with generation of $e^+/e^-$.\newline
390 2 Pair production without generation of $e^+/e^-$.\\
392 22 PHOT & 1 & Photoelectric effect. The interacting photon is
394 0 No photo-electric effect. \newline
395 1 Photo-electric effect with generation of $e^-$.\newline
396 2 Photo-electric effect without generation of $e^-$.\\
398 23 RAYL & 1 & Rayliegh effect\cite{Rayligh}. The interacting
399 $\gamma$ is not stopped.\newline
400 0 No Raylieght effect.\newline
403 24 STRA & 0 & Turns on the collision sampling method to simulate
404 energy loss in thin materials, particularly gasses.\newline
405 0 Collision sampling is off.\newline
406 1 Collision sampling is on. \\
408 PFIS & 0 & Nuclear fission induced by a photon The photon stops.\newline
409 0 No photo-fission.\newline
410 1 Photo-fission with generation of secondaries.\newline
411 2 Photo-fission without generation of secondaries.\\
413 MUNU & 1 & Muon-nucleus interactions. The muon is not stopped.\newline
414 0 No muon-nucleus interactions.\newline
415 1 Muon-nucleus interactions with generation of
417 2 Muon-nucleus interactions without generation of secondaries.\\
419 CKOV & 1 & Light absorption. This process is the absorption of light
420 photons in dielectric materials. It is turned on by default
421 when the generation of $\check{C}$erenkov\cite{Cerenkov}
422 light is requested (in GEANT manual it is LABS).\newline
423 0 No absorption of photons.\newline
424 1 Absorption of photons with possible detection.\\
426 SYNC & 0 & Synchrotron radiation in magnetic fields.\newline
427 0 Synchrotron radiation is not simulated.\newline
428 1 Synchrotron photon are generated, at the end of the
429 tracking step.\newline
430 2 Photons are not generated, the energy is deposited
432 3 Synchrotron photons are generated, distributed along the
433 curved path of their particle. \\
435 \label{table:GEANT3PhysicsFlags}
438 \begin{longtable}{p{0.15\textwidth}p{0.2\textwidth}p{0.65\textwidth}}
439 \multicolumn{3}{l}{Table \ref{table:GEANT3PhysicsLimits}} \\
441 Parameter & \small ALICE Default value & Description \\ \hline
443 \multicolumn{3}{l}{\emph{Table \ref{table:GEANT3PhysicsLimits}
446 Parameter & ALICE Default value & Description \\
450 \multicolumn{3}{r}{\emph{Table \ref{table:GEANT3PhysicsLimits}
451 continued on next page.}}
454 \caption{\label{table:GEANT3PhysicsLimits}GEANT3 physics process limits.
455 These ``cuts'' can be set on a
456 material by material basis. The ALICE Default values are
457 set in the \texttt{Config.C} file uses the
458 \texttt{gMC->SetCuts}
459 function. The setting of these specific flags for any
460 specific material is done in
461 \texttt{\$ALICE\_ROOT/data/galice.cuts}
462 file. The number on the left of the cut name is the
463 column in the \texttt{galice.cuts} file that this cut
464 is expected to be found. This information comes from the
465 GEANT3 documentation ZZZZ010-2 \cite{GEANT3:documentatoin}
470 3 CUTGAM & $1.\times 10^{-3}$ GeV & Threshold for gamma transport.\\
472 4 CUTELE & $1.\times 10^{-3}$ GeV & Threshold for electron and positron
475 5 CUTNEU & $1.\times 10^{-3}$ GeV & Threshold for neutral hadron
478 6 CUTHAD & $1.\times 10^{-3}$ GeV & Threshold for charged hadron
481 7 CUTMUO & $1.\times 10^{-3}$ GeV & Threshold for muon transport.\\
483 8 BCUTE & $1.\times 10^{-3}$ GeV & Threshold for photons produced by
484 electron bremsstrahlung.\\
486 9 BCUTM & $1.\times 10^{-3}$ GeV & Threshold for photons produced by
487 muon bremsstrahlung.\\
489 10 DCUTE & $1.\times 10^{-3}$ GeV & Threshold for electrons produced by
490 electron $\delta$-rays.\\
492 11 DCUTM & $1.\times 10^{-3}$ GeV & Threshold for electrons produced by
493 muon or hadron $\delta$-rays.\\
495 12 PPCUTM & $1.\times 10^{-3}$ GeV & Threshold for $e^{\pm}$ direct pair
496 production by muons.\\
498 TOFMAX & $1.\times 10^{10}$ sec & Threshold on time of flight counted
499 from primary interactions time.\\
501 \label{table:GEANT3PhysicsLimits}
504 \paragraph{GEANT4 Switches and Settings}
506 \paragraph{Fluka Switches and Settings}
509 \begin{thebibliography}{99}
510 \bibitem{GEANT3:documentatoin} CERN Program Library.
511 \textsl{GEANT Detector description and Simulation Tool},
512 CERN Program Library Long writeup W5013, October 1994
513 \bibitem{GEANT4} GEANT4 Working Group, ``User Documentation'',
514 Accessed Feb. 4 2013.
515 http://geant4.cern.ch/support/suerdocumentatoin.shtml
516 \bibitem{Fluka} FLUKA Team, ``FLUKA'', Access Feb. 4 2013,
517 http://www.fluka.org/fluka.php?id=man\_onl
518 \bibitem{LandauFluct} \textsl{GEANT Detector description and Simulation Tool},
520 L. Landau. \textsl{On the Energy Loss of Fast Particles
521 by Ionisation}, J. Phys. 8:201, 1944. and
522 K. S. K$\ddot{o}$lbig and B. Schorr. \textsl{Asymptotic
523 expansion for the Landau density and distribution
524 functions}, Comp. Phys. Comm., 32:121,1984
525 \bibitem{LadauVavilov}\textsl{GEANT Detector description and Simulation Tool},
527 P. V. Valilov. \textsl{Ionisation losses of high
528 energy heavy particles}, Soviet Physics
530 \bibitem{Moliere}\textsl{GEANT Detector description and Simulation Tool},
532 G. Z. Moliere \textsl{Theorie de Streuung schneller
533 geladener Teilchen I: Enzelstreuung am abgerschmitten
534 Coulomb-Feld},Z. Naturforsch., 2a:133, 1947 and
535 G. Z. Moliere, \textsl{Teeorie der Steuung schneller
536 geladerner Teichen II: Merfach- und Vielfachstreuung}
537 Z. Naturforsh., 3a:87, 1948 and
538 W. T. Scott. Rev. Mod. Phys. 35:231 1963
539 \bibitem{Rossi}\textsl{GEANT Detector description and Simulation Tool},
541 R. Rossi, Prentice-Hall, Englewoood Clifgs, 1962
542 R. Rossi, and K. Greisen, Rev. Mod. Physics, 13-240, 1942
543 \bibitem{Rayliegh} \textsl{GEANT Detector description and Simulation Tool},
545 W. R. Nelson, H. Hiayama, and D. W. O. Rogers. Technical
546 Report 265, SLAC, 1985
547 \bibitem{Cerenkov} \textsl{GEANT Detector description and Simulation Tool},
549 J. D. Jackson \textsl{Classical Electrodyanamics},
550 J. Wiley et Sons, Inc. New York, 1975
551 \end{thebibliography}