1: \documentclass[tbtags,12pt,a4paper]{article}

2: \usepackage{a4wide}

3: \usepackage{fancyhdr}

4: \usepackage{graphicx}

5: \usepackage{float}

6: \usepackage{enumerate}

7: \usepackage{amsmath}

8: \usepackage{amssymb}

9: \usepackage{amsfonts}

10: \usepackage{amsthm}

11: \usepackage{amssymb}

12: \usepackage{mathrsfs}

13: %\usepackage{subeqn}

14: \usepackage{equation}

15: 

16: \begin{document}

17: \renewcommand{\thefootnote}{\fnsymbol{footnote}}

18: \def\gsim{\:\raisebox{-0.5ex}{$\stackrel{\textstyle>}{\sim}$}\:}

19: \def\lsim{\:\raisebox{-0.5ex}{$\stackrel{\textstyle<}{\sim}$}\:}

20: 

21: \begin{titlepage}

22: 

23: \begin{flushright}

24: RIKEN--TH 86 \\[-0.1cm]

25: October 2006\\[5mm]

26: \end{flushright}

27: 

28: \vspace{0.5cm}

29: 

30: \begin{center}

31: {\Large \bf QCD Effects in the Decays of TeV Black Holes}\\[1.cm]

32:    {\large Christian Alig$^{1}$, Manuel Drees$^1$, and Kin-ya Oda$^2$}

33: \end{center}

34: 

35: \vskip 0.5cm

36: 

37: {\small

38: \begin{center}

39: 

40: $^1$ {\it Physikalisches Institut, Universit\"at Bonn, Nussallee 12, D53115

41:           Bonn, Germany} \\[2mm]

42: $^2$ {\it Theoretical Physics Laboratory, RIKEN, Saitama 351-0198, Japan}

43: 

44: \end{center}

45: }

46: 

47: \vspace{1.cm}

48: \begin{abstract}

49: 

50: \noindent In models with ``large'' and/or warped extra dimensions, the

51: higher--dimensional Planck scale may be as low as a TeV. In that case black

52: holes with masses of a few TeV are expected to be produced copiously in

53: multi--TeV collisions, in particular at the LHC. These black holes decay

54: through Hawking radiation into typically ${\cal O}(20)$ Standard Model

55: particles. Most of these particles would be strongly interacting. Naively this

56: would lead to a final state containing 10 or so hadronic jets. However, it

57: has been argued that the density of strongly interacting particles would be so

58: large that they thermalize, forming a ``chromosphere'' rather than

59: well--defined jets. In order to investigate this, we perform a QCD simulation

60: which includes parton--parton scattering in addition to parton showering. We

61: find the effects of parton scattering to remain small for all cases we

62: studied, leading to the conclusion that the decays of black holes with masses

63: within the reach of the LHC will {\em not} lead to the formation of

64: chromospheres.

65: 

66: \end{abstract}

67: \vskip 0.5cm

68: 

69: \end{titlepage}

70: 

71: \section{Introduction}

72: 

73: The Standard Model (SM) of particle physics is very successful in reproducing

74: experimental data. However, from the theoretical point of view it has several

75: unsatisfactory features. Its perhaps biggest problem stems from the very large

76: hierarchy between the electroweak mass scale $M_{\rm weak} \sim 100$ GeV and

77: the (reduced) Planck mass $M_{\rm P} \simeq 2.4 \cdot 10^{18}$ GeV. The origin

78: of this hierarchy has no explanation within the SM; worse, it tends to be

79: destroyed by quantum corrections.

80: 

81: In the late 1990's it has been suggested \cite{Arkani-Hamed:1998rs,

82:   Randall:1999ee} that this problem could be solved by introducing additional

83: spatial dimensions. They could be relatively large if only gravity is allowed

84: to propagate in them, i.e. if all (other) SM fields are bound to a brane with

85: three spatial dimensions. The strength of the gravitational interactions at

86: lengths larger than the radii of these additional dimensions would then be

87: diluted by a factor which is essentially given by the ratio of the volume of

88: these extra dimensions and the appropriate power of the higher--dimensional

89: Planck length. This ratio can be very large even if the higher--dimensional

90: Planck mass $M_D$ is not far above $M_{\rm weak}$. If this idea is correct,

91: the very large measured value of $M_{\rm P}$ is an artefact of using a

92: 3--dimensional description of a world that actually has $3+n$ spatial

93: dimensions.

94: 

95: This scenario leads to dramatic predictions for collisions of pointlike

96: particles at high center--of--mass (cms) energy. Collisions at energies around

97: $M_D$ would likely be dominated by the exchange of gravitons

98: \cite{'tHooft:1987rb}, rather than by exchange of SM gauge bosons. Collisions

99: at energies exceeding $M_D$ could lead to the formation of black holes

100: \cite{banks, Giddings:2001bu, Dimopoulos:2001hw, Eardley:2002re,

101:   Solodukhin:2002ui, Hsu:2002bd, Yoshino:2002tx, Yoshino:2005hi}, with cross

102: section being given by the square of the (generalized) Schwarzschild radius.

103: Since this radius actually increases with increasing cms energy, this would

104: lead to ``the end of short--distance physics'' \cite{Giddings:2001bu}.

105: 

106: These black holes with masses of a few TeV should decay very quickly via the

107: higher--dimension analogue \cite{Myers:1986un} of Hawking radiation

108: \cite{Hawking:1974sw}, into final states on the brane consisting mostly of SM

109: degrees of freedom \cite{Emparan:2000rs}. Since the Hawking temperature of

110: black holes that can be produced at the LHC is ${\cal O}(100 \ {\rm GeV})$,

111: its decay should produce roughly ${\cal O}(20)$ particles with average energy

112: $\sim 3$ times the Hawking temperature. This decay would appear to be

113: instantaneous to LHC experiments, since the typical lifetime of such a black

114: hole is only ${\cal O}(10^{-27} \ {\rm s})$.

115: 

116: Since these black holes decay into all SM degrees of freedom with

117: approximately equal probability, most final state particles would be strongly

118: interacting quarks or gluons. This leads to the expectation that a typical

119: black hole event at the LHC would contain 10 or so jets plus a few (charged or

120: neutral) leptons and/or photons, each with typical energy of several hundred

121: GeV. This is obviously a very dramatic signature, which should be easy to

122: detect.

123: 

124: However, it has been argued by Anchordoqui and Goldberg

125: \cite{Anchordoqui:2002cp} that the density of strongly interacting partons

126: just after the decay of the black hole would be so high that they would

127: frequently interact with each other, leading to the formation of a (more or

128: less) thermalized ``chromosphere'', i.e. a (nearly) spherical shell of

129: thousands of particles with rather small energies. This final state would also

130: be very easy to detect. However, in this case the final state would be

131: characterized almost uniquely by the mass of the black hole. More detailed

132: investigations of the primary decay spectrum, which could give information

133: about the number of additional dimensions as well as the spin of the decaying

134: black hole, could then only be performed with the (rather few) primary charged

135: leptons and photons.

136: 

137: It is therefore of some importance to decide whether the assertion of

138: ref.\cite{Anchordoqui:2002cp} is in fact correct. In this paper we report

139: results of a Monte Carlo simulation of the QCD effects relevant for the decay

140: of TeV black holes, including both parton showering and partonic collisions.

141: This simulation has to keep track of the space--time evolution of the black

142: hole decay products. In contrast, the usual shower codes only keep track of

143: the virtuality of the partons in the shower, but treat the shower itself as

144: instantaneous. This is quite adequate for most applications, since a parton

145: shower only lasts $10^{-23}$ seconds or so, corresponding to a spatial

146: extension of a few Fermi, many orders of magnitude below the resolution of any

147: conceivable detector. However, if a chromosphere forms at all, it should do so

148: during those $10^{-23}$ seconds. A careful treatment of the spatio--temporal

149: evolution of the shower is therefore mandatory for us.

150: 

151: We find that the effects of parton--parton scatterings after black hole decay

152: are essentially negligible for a black hole of 5 TeV, and are quite small even

153: for 10 TeV black holes. Even in the latter case, most partons do not scatter.

154: Moreover, most of these (relatively rare) interactions are rather soft, i.e.

155: they do not change the energies, trajectories or virtualities of the

156: participating partons very much. We therefore conclude that the decays of

157: black holes that might be produced at the LHC will {\em not} lead to the

158: formation of a chromosphere. Black hole event generators that ignore

159: interactions between black hole decay products \cite{charybdis,catfish} only

160: make a small mistake.

161: 

162: The remainder of this note is organized as follows. In the next Section we

163: describe the production of TeV black holes, and their subsequent decay through

164: Hawking radiation, in slightly more detail. This determines the initial

165: set--up of the partonic system, which later may or may not develop into a

166: chromosphere. In Sec.~3 we summarize the argument in favor of chromosphere

167: formation, closely following ref.\cite{Anchordoqui:2002cp}; we also point out

168: some weaknesses in this argument. Sec.~4 is devoted to a description of the

169: simulation program we wrote. In Sec.~5 we present numerical results for the

170: angular correlation between pairs of charged particles, for the overall energy

171: flow of the hadronic black hole decay products, and for the microscopic

172: structure of these events. Finally, Sec.~6 contains a brief summary and

173: outlook.

174: 

175: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

176: \section{Black hole production and decay}

177: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

178: 

179: Lower bounds on the classical gravitational production cross section of black

180: holes are obtained in~\cite{Eardley:2002re,Yoshino:2002tx,Yoshino:2005hi}.

181: Quantum corrections are estimated in~\cite{Solodukhin:2002ui,Hsu:2002bd}.

182: The implication of the correspondence principle for black holes and strings is

183: considered in~\cite{Dimopoulos:2001qe}. All these results suggest that the

184: black hole production cross section grows geometrically above the

185: (higher--dimensional) Planck scale:

186: %

187: \begin{align} \label{cs}

188: \sigma(\hat{s}) &\simeq \pi r_h(M)^2 \propto \hat{s}^{1/(1+n)} \, .

189: \end{align}

190: %

191: Here, $n$ is the number of additional spatial dimensions, $\hat{s}$ is the

192: partonic cms energy, and $r_h$ is the horizon radius of the

193: $D=4+n$~dimensional Schwarz\-schild black hole with mass

194: $M=\sqrt{\hat{s}}$:\footnote{We use the notation of \cite{Anchordoqui:2002cp},

195:   where $M_D$ stands for the reduced higher--dimensional Planck mass.}

196: %

197: \begin{equation} \label{radius}

198: r_h(M) = {1 \over M_D} \left[ \frac {M} {M_D} \frac {2^n \pi^{(n-3)/2} 

199: \Gamma\left({3+n\over2}\right) } {n+2} \right]^{1/(1+n)} \, .

200: \end{equation}

201: %

202: Eq.(\ref{cs}) should hold as long as $r_h$ is small compared to the size of

203: the additional dimensions, which is true for all cases of interest to LHC

204: experiments. 

205: 

206: Astrophysical processes lead to the very strong lower bound $M_D \gg 10$ TeV

207: for $n \leq 3$ \cite{raffelt}. For larger $n$ the lower bound on $M_D$ comes

208: from searches for the production of gravitons (including their Kaluza--Klein

209: towers) at colliders, yielding $M_D \geq 0.65$ TeV for $n=6$ \cite{pdg}.  It

210: has been argued \cite{anchor3} that the non--observation of black holes

211: produced by very energetic cosmic ray neutrinos yields the slightly stronger

212: bound $M_D \gsim 1$ TeV for $n \geq 5$. However, this bound relies on

213: assumptions on the flux of very energetic cosmic neutrinos. We will see

214: shortly that chromosphere formation is most likely for the smallest allowed

215: value of $M_D$. To be conservative, we will therefore present numerical

216: results for $n=6$ and $M_D = 0.65$ TeV. In that case according to

217: Eqs.(\ref{cs}), (\ref{radius}) the LHC operating at a proton--proton cms

218: energy of 14 TeV yields cross sections in excess of 200 pb (10 fb) for $M > 5

219: \ (10)$ TeV~\cite{Giddings:2001bu}. In other words, the LHC will be a

220: veritable black hole factory if a TeV scale gravity scenario is employed by

221: nature~\cite{Giddings:2001bu,Dimopoulos:2001hw}.\footnote{QCD initial state

222:   radiation is, as usual, taken into account by using scale--dependent parton

223:   distribution functions, the relevant momentum scale being set by $1/r_h$

224:   \cite{Giddings:2001bu}. Numerical simulations indicate \cite{Yoshino:2002tx,

225:     Yoshino:2005hi} that several tens of percent of the cms energy of the

226:   colliding partons may escape in form of gravitational radiation. The same

227:   calculations show that Eqs.(\ref{cs}), (\ref{radius}) underestimate the

228:   cross section for black hole production for fixed $\hat{s}$. Nevertheless

229:   the energy ``lost'' in gravitational radiation implies $M < \sqrt{\hat{s}}$,

230:   in which case the cross section for the production of black holes with a

231:   given mass at the LHC might be more than three orders of magnitude smaller

232:   than indicated by Eqs.(\ref{cs}), (\ref{radius}) with $\sqrt{\hat{s}} = M$

233:   \cite{anchor2}. Finally, if a ``generalized uncertainty principle'' imposes

234:   a lower bound on physical lengths of order $1/M_D$, the cross section for

235:   producing black holes with mass $M \gg M_D$ at the LHC would also be reduced

236:   by several orders of magnitude \cite{hoss1}.}

237:   

238: We note that the black hole is generically produced with a sizable angular

239: momentum~\cite{Ida1}. However, here we only consider the case of vanishing

240: angular momentum. While the spin of the black hole would affect the details of

241: its decay spectrum during the early ``spin--down'' phase \cite{Ida1, kanti,

242:   Ida:2006tf}, these details are not likely to significantly change the

243: importance of the QCD processes which are the main focus of our analysis.

244: 

245: Once produced, the black hole radiates off its mass~$M$ via Hawking

246: radiation~\cite{Hawking:1974sw} mainly into the brane--localized standard model

247: particles~\cite{Emparan:2000rs}.\footnote{Possible enhancement effects of bulk

248:   graviton emission, especially for highly rotating black holes, have been

249:   discussed in \cite{enhance}.} The number spectrum for the emission of a

250: spin$-s$ particle with energy $\omega$ per unit time is

251: %

252: \begin{align} \label{spectrum}

253: {d \dot{N}_s\over d\omega} &= {1\over2\pi} {\Gamma_s\over e^{\omega/T} -

254:   (-1)^{2s}}\, . 

255: \end{align}

256: %

257: Here the Hawking temperature $T$ is given by

258: %

259: \begin{align} \label{temp}

260: T &= {1+n\over 4\pi r_h}\, .

261: \end{align}

262: %

263: The ``greybody factor'' $\Gamma_s(E)$ in Eq.(\ref{spectrum}) is defined to be

264: the absorption rate of the incoming flux with energy $E$ at spatial infinity:

265: %

266: \begin{align}  \label{gb}

267: \Gamma_s &= {\dot{N}_\text{in} - \dot{N}_\text{out} \over \dot{N}_\text{in}},

268: \end{align}

269: %

270: when the purely in--going boundary condition is put at the black hole horizon.

271: Physically $\Gamma_s$ is, in the ``time-reversed'' sense, the proportion of

272: the radiation that passes through the gravitational potential well from the

273: horizon towards spatial infinity.

274: 

275: \begin{figure}[H] \label{specfig}

276: \begin{center}

277: \includegraphics[scale=2]{number_spectra.eps}

278: \end{center}

279: \caption{The energy spectrum of gluons (upper, black) and quarks (lower, grey)

280:   emitted in black hole decay for $n=6$ additional dimensions. $\omega$ is the

281:   energy of the emitted particle, and the Schwarzschild radius $r_h$ is given

282:   by Eq.(\ref{radius}).}

283: \end{figure}

284: 

285: \setcounter{footnote}{0}

286: Numerical results \cite{Ida:2006tf} for the spectra of quarks and gluons

287: produced in the decay of a black hole are shown in Fig.~1. In principle

288: Eqs.(\ref{spectrum}) and (\ref{temp}) predict that the spectrum of black hole

289: decay products changes with time: as more energy is radiated off, the mass of

290: the black hole becomes smaller. Eq.(\ref{radius}) shows that this also reduces

291: its radius, which according to Eq.(\ref{temp}) increases its

292: temperature.\footnote{Since Fig.~1 effectively shows the spectrum in units of

293:   the inverse black hole radius, it includes all the information required for

294:   including this effect in the simulation. However, these semi--classical

295:   results only hold for $M \gg \omega$; they may therefore not describe the

296:   late stages of black hole decay adequately. In fact, it has been argued

297:   \cite{remnant} that a black hole will not evaporate completely, but leave

298:   behind a stable or at least long--lived, and possibly charged, remnant with

299:   mass $\sim M_D$. We do not consider this possibility in our work.} For

300: simplicity we employ the ``sudden decay approximation''

301: \cite{Dimopoulos:2001hw} where the entire decay spectrum is calculated using a

302: fixed black hole mass.

303: 

304: All degrees of freedom with a given spin are emitted with equal probability in

305: black hole decay; however, Fig.~1 shows that the greybody factors do depend on

306: the spin. In order to include this effect, we define the integrals over the

307: decay spectra

308: %

309: \begin{equation} \label{is}

310: I_s = \int_0^\infty \frac {d \dot{N}_s} {d \omega} d z \, ,

311: \end{equation}

312: %

313: where $z = \omega r_h$ is the dimensionless variable shown in Fig.~1.

314: Numerically, $I_{1/2} \simeq 0.115, \, I_1 = 0.155$. In our simulation we are

315: primarily interested in the question whether quarks and gluons produced in

316: black hole decay will thermalize. Since scattering cross sections for heavy

317: quarks are smaller than those for massless quarks, we conservatively assume

318: that only $u,d,s,c$ quarks and gluons are emitted in black hole decays. The

319: relative abundance of gluons is then on average given by

320: %

321: \begin{equation} \label{gluons}

322: \frac {\langle N_g \rangle} {\langle N_g \rangle + \langle N_q \rangle} =

323: \frac {8 I_1} {8 I_1 + 24 I_{1/2}} \, . 

324: \end{equation}

325: %

326: An analogous argument shows that about 75\% of the black hole mass will be

327: radiated into quarks and gluons.\footnote{Massive gauge and Higgs bosons

328:   produced in black hole decay frequently also decay into quarks. However,

329:   these bosons have much longer lifetimes than the black hole itself. The

330:   quarks produced in their decays therefore come too late to contribute to the

331:   formation of a chromosphere, although they might get trapped in the

332:   chromosphere \cite{Anchordoqui:2002cp} should it indeed form.}

333: 

334: The total (average) number of particles produced in the decay of a single

335: black hole can be estimated from the integrals

336: %

337: \begin{equation} \label{its}

338: \tilde I_s = \int_0^\infty z \frac {d \dot{N}_s} {d \omega} d z \, ;

339: \end{equation}

340: %

341: numerically, $\tilde I_{1/2} = 0.181, \, \tilde I_1 = 0.231$. The average

342: energies (in units of $r_h^{-1}$) of the produced particles are given by the

343: ratios $\tilde I_s / I_s$. Numerically, the average energy of a quark or gluon

344: is $\langle E_q \rangle = 313 \ (283)$ GeV and $\langle E_g \rangle = 296 \ 

345: (268)$ GeV for $M = 5 \ (10)$ TeV. Using the fact that about 75\% of the total

346: energy goes into hadrons, this gives average parton multiplicities

347: %

348: \begin{eqnarray} \label{nav}

349: \langle N_q \rangle &=& 8.4 \ (18.6) \nonumber \\

350: \langle N_g \rangle &=& 3.8 \ (8.3) \, ,

351: \end{eqnarray}

352: %

353: for $M = 5 \ (10)$ TeV, where $N_q$ counts both quarks and antiquarks.

354: Choosing a larger higher--dimensional Planck scale $M_D$ would lead to higher

355: Hawking temperature $T$, and hence to fewer, more energetic, primary black

356: hole decay products; this would obviously reduce the probability of

357: parton--parton scattering.

358: 

359: Finally, Eq.(\ref{spectrum}) also allows to compute the lifetime of the black

360: hole. To that end, one computes the time derivative of the mass of the black

361: hole, 

362: %

363: \begin{equation} \label{dmdt}

364: \frac {dM} {dt} = \int_0^\infty \omega \frac {d \dot{N} } {d \omega} d

365: \omega\, .

366: \end{equation}

367: %

368: If we set $t=0$ for the time of black hole production, its lifetime $\tau_{\rm

369:   bh}$ is defined by $M(\tau_{\rm bh})$ = 0. Eq.(\ref{spectrum}) shows that

370: the rhs of Eq.(\ref{dmdt}) is $\propto r_h^{-2}$. The use of Eq.(\ref{radius})

371: then leads to \cite{Giddings:2001bu}

372: %

373: \begin{equation} \label{life}

374: \tau_{\rm bh} = \Gamma_{\rm bh}^{-1} = \frac {C} {M_D} \left(\frac {M} {M_D}

375: \right)^{(n+3)/(n+1)}\, .

376: \end{equation}

377: %

378: The coefficient $C$ can be computed by numerically integrating the spectra

379: shown in Fig.~1, leading to $C \simeq 0.2$. Eq.(\ref{life}) then

380: gives $\Gamma_{\rm bh} = 240\ (100)$ GeV for $M= 5 \ (10)$ TeV and $M_D = 0.65$

381: TeV, corresponding to lifetimes of order $10^{-27}$ seconds.

382: 

383: Having presented the relevant properties of the partonic final state that

384: emerges from black hole decay, we are now ready to discuss whether

385: interactions between these partons might lead to formation of a chromosphere.

386: 

387: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

388: \section{Arguments for and against chromosphere formation}

389: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

390: 

391: The argument by Anchordoqui and Goldberg \cite{Anchordoqui:2002cp} starts from

392: the observation that the parton number density $n$ after black hole decay is

393: quite high: roughly ${\cal O}(10)$ partons are distributed over a sphere with

394: radius $\sim c \tau_{bh}$. They estimate the rate $\Gamma$ of bremsstrahlung

395: reactions (which increase the total number of partons) as $\Gamma = n c

396: \sigma_b$, with 

397: %

398: \begin{equation} \label{brems}

399: \sigma_b \simeq \frac {8 \alpha_s^3} {\Lambda^2} \ln \left( \frac {2Q}

400:   {\Lambda} \right) \, ,

401: \end{equation}

402: %

403: where $\alpha_s$ is the strong coupling constant and $Q$ the initial energy of

404: the parton in the rest frame of the black hole; the energy scale $\Lambda$ is

405: estimated as the inverse of the radius of the expanding shell of partons. This

406: leads to a total interaction rate per parton \cite{Anchordoqui:2002cp}

407: %

408: \begin{equation} \label{intrate}

409: {\cal N}_{\rm int} \simeq 0.15 \frac {N_{q,{\rm init}}} {10} \left( \frac

410:   {\alpha_s(Q_{\rm min}) } {0.2} \right)^3 \ln \left( \frac {2 Q} {Q_{\rm min}}

411:   \right) \ln \left( \frac {\Gamma_{\rm bh}} {Q_{\rm min}} \right) \, .

412: \end{equation}

413: %

414: For $Q \simeq 400$ GeV and initial parton number $N_{q,{\rm init}}=10$

415: Eq.(\ref{intrate}) predicts about 3 interactions per parton even with minimal

416: momentum transfer $Q_{\rm min} = 9$ GeV; this increases to $\sim 30$

417: interactions for $Q_{\rm min} =1.8$ GeV (the mass of the $\tau$

418: lepton). Anchordoqui and Goldberg argue that partons that interact so

419: frequently must thermalize, leading to the formation of an expanding shell of

420: particles with approximately thermal energy distribution.\footnote{In

421:   contrast, the authors of ref.\cite{Giddings:2001bu} state that the shell of

422:   black hole decay products is too thin to thermalize; however, no

423:   quantitative estimate of the effects of interactions between these decay

424:   products is given there.}

425: 

426: This argument has several weaknesses. To begin with, most QCD interactions are

427: rather soft, as seen by the factor $1/\Lambda^2$ in Eq.(\ref{brems}).

428: Similarly, most emitted gluons will be soft and/or collinear; this leads to

429: the logarithmic factor in the cross section (\ref{brems}). In other words, a

430: ``typical'' interaction may not change the energies and momenta of the

431: participating partons very much, and may only lead to the emission of a soft

432: and/or collinear gluon. Such interactions would {\em not} impede the formation

433: of well--defined hadronic jets. In fact, during a QCD parton cascade, many

434: gluons will in any case be emitted from the partons produced in a given

435: ``hard'' process; most of these gluons will also be soft and/or collinear.

436: This gives rise to finite widths and masses of hadronic jets, but does not

437: destroy them.

438: 

439: Secondly, Anchordoqui and Goldberg seem to have overlooked the fact that a

440: scattering reaction takes a finite time: according to the uncertainty

441: principle, the time at which a reaction with energy exchange of ${\cal

442:   O}(\Lambda)$ occurs can only be determined with an intrinsic uncertainty of

443: order ${\cal O}(1/\Lambda)$. Most reactions have $\Lambda \sim r^{-1}$

444: \cite{Anchordoqui:2002cp}, $r$ being the radius of the expanding shell of

445: particles. Loosely speaking, a particle ``has time'' for only one such

446: reaction while traveling a distance ${\cal O}(r)$.

447: 

448: The formation of a chromosphere seems unlikely on purely phenomenological

449: grounds. Note that the number of interactions per parton is proportional to

450: the number of initial partons. If ten initial partons lead to a chromosphere,

451: five or six initial partons should at least show significant effects from

452: these interactions. The observation of events with six well--defined jets has

453: been reported by both the UA2 \cite{ua2} and CDF \cite{cdf} collaborations.

454: CDF finds fair agreement between observations and QCD parton shower

455: simulations (based on leading--order matrix elements). They demand that the

456: total invariant mass of the 6--jet system exceed 520 GeV, while each jet

457: should have a transverse momentum of at least 20 GeV. Using $N_{\rm init} =

458: 6$, $Q = 30$ GeV and $\Gamma_{\rm bh} \longrightarrow M_{6-{\rm jet}} = 500$

459: GeV setting the scale for the initial hard reaction, Eq.(\ref{intrate})

460: predicts ${\cal N} \simeq 0.7 \ (9.5)$ interactions per parton for $Q_{\rm

461:   min} = 9 \ (2)$ GeV. We find it difficult to believe that such high

462: interaction rates would leave no imprint in the properties of the observed

463: events, given that 3 (30) interactions are supposed to lead to nearly complete

464: thermalization. 

465: 

466: Even if a nearly thermal chromosphere does not form, parton--parton scattering

467: might still have significant impact on the hadronic final state from black

468: hole decay. Given the complexity of QCD processes even in the absence of

469: parton--parton scattering, a quantitative investigation of their effect can

470: only be performed with the help of a QCD simulation program. This is the topic

471: of the next Section.

472: 

473: 

474: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

475: \section{Simulation}

476: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

477: 

478: We saw in the previous Section that a quantitative analysis of the effects of

479: parton--parton scattering is only possible if we also treat the QCD parton

480: showers that occur whenever a large four--momentum is transmitted to strongly

481: interacting particles. This implies that we need to follow the

482: spatio--temporal development of these QCD showers. In this Section we first

483: outline the general philosophy of our approach; a more detailed description

484: of the various stages treated by our code will be given in the subsequent

485: Subsections. 

486: 

487: Numerical codes that simulate parton showers are probabilistic, i.e. they

488: operate with squared amplitudes. Quantum mechanical interference effects can

489: therefore only be treated approximately (e.g. through angular ordering

490: \cite{order}, whereby subsequent gluons are emitted at smaller and smaller

491: angles.) The basic idea is that partons emerging from a hard reaction

492: (scattering or decay) initially have time--like virtuality, which is reduced

493: by ``branching off'' additional partons. This can be justified by the

494: observation that in a complete calculation of the relevant Feynman diagrams,

495: final state partons emitting additional partons indeed have to have time--like

496: virtualities. The beauty of such showering algorithms is that they exploit QCD

497: factorization theorems to sum such higher order processes to all orders in

498: perturbation theory, albeit (usually) only with leading logarithmic accuracy.

499: 

500: Most of these codes follow the ``evolution'' of this shower not in time, but

501: in an energy variable, which in the simplest case is given by the virtuality

502: of the partons in the shower; since we are dealing with final--state showers,

503: the partons in question have time--like momenta, i.e. the shower has the same

504: kinematics as a cascade of two--body decays. (In this analogy, the lifetime of

505: the decaying particles would be given by the inverse virtuality of the

506: particles in the shower.)

507: 

508: According to quantum mechanics, one cannot simultaneously determine this

509: shower energy scale and the (proper) time of a branching. Unfortunately for

510: our application we need to do precisely that. A certain additional abuse of

511: the principles of quantum mechanics is therefore inevitable. We do this by

512: identifying the {\em uncertainty} in time, as determined by the uncertainty

513: principle, with the actual {\em duration} of a given process. We use this

514: identification both for the branching and for parton--parton scattering; in

515: the former case, the time is given by the inverse of the virtuality of the

516: branching parton, whereas the time needed for a scattering is given by the

517: (space--like) virtuality of the parton exchanged in this scattering reaction.

518: Note that neither two branching steps, nor two scatterings, involving the same

519: partons can occur at the same time. In the absence of scattering, the

520: evolution in time would therefore strictly match the evolution towards smaller

521: virtualities. Moreover, between branching or scattering events, the partons

522: are taken to propagate along their classical paths. Note that this evolution

523: is nonetheless probabilistic, since after each branching or scattering the

524: 4--momenta of the outgoing partons are chosen randomly, with distributions

525: determined by perturbative QCD (and subject to energy--momentum conservation).

526: 

527: We only aim at leading logarithmic accuracy. This means that we only use

528: leading order cross sections, and only include $2 \rightarrow 2$ scattering

529: processes; later gluon emission, which is treated explicitly in the estimates

530: of interaction rates in ref.\cite{Anchordoqui:2002cp}, is taken into account

531: by the subsequent shower evolution. In fact, including $2 \rightarrow 3$

532: processes in the scattering reactions would lead to double counting.

533: Scattering reactions can nevertheless increase the particle multiplicities,

534: since they can {\em increase} the virtuality of the participating particles;

535: in contrast, each branching reduces this virtuality. Moreover, the scattering

536: can change the 3--momenta of the particles, thereby potentially destroying the

537: jet structure. As mentioned in the previous Section, one needs large

538: scattering angles in order to establish a chromosphere from an initially small

539: number of very energetic partons.

540: 

541: As usual, we treat showering and hard scattering independently, i.e. we apply

542: QCD factorization. This requires that the scattering indeed be sufficiently

543: hard; that is, the absolute value of the four--momentum exchanged in the

544: scattering reaction must be (much) larger than the virtualities of the partons

545: in both the initial and final state. However, as already noted, the partons in

546: the final state may be more off--shell than those in the initial state. Note

547: also that we should not include initial state showering in our approach, since

548: each ``initial state'' of a scattering reaction is part of the extended final

549: state shower that follows the Hawking evaporation of the black hole; including

550: initial state radiation would therefore also lead to double counting.

551: 

552: Our simulation starts with a rather small number of energetic, and far

553: off--shell, partons, using the results of Section~2. It then uses small time

554: steps to follow the parton shower and/or scattering of all partons. This phase

555: ends when all partons are (nearly) on--shell and far apart from each

556: other, so that neither further branching nor further scattering is possible.

557: When the particles virtualities reach the QCD scale $\Lambda_{QCD}$,

558: hadronization will take place and the data of the final particles will be

559: stored for statistical analysis.

560: 

561: The simulation code has been written nearly from scratch in C++, although the

562: global structure is based on the VNI 4.12b Monte Carlo simulation

563: \cite{Geiger:1998fq, Bass:1999pv} using its particle record and a selection of

564: modified routines from it. The VNI particle record uses the ``Les Houches''

565: format, extended to hold information necessary for the full space--time

566: evolution of the partons. VNI 4.12b was originally written in order to

567: simulate ultra--relativistic heavy--ion collisions; in that case, multiple

568: partonic scattering reactions are certainly important, and had to be modeled

569: carefully, making this code a good starting point for our work. In contrast,

570: scattering in the final state is thought to be unimportant for hard reactions

571: involving $e^\pm$ and/or $p/\bar p$ in the initial state.

572: 

573: We also use some routines from the PYTHIA Monte Carlo simulation

574: \cite{Sjostrand:2003wg}. For numerical integration and a simulation grade

575: random number generator the GNU Scientific Library \cite{gnu} for C/C++ was

576: used.

577: 

578: \subsection{Initial setup}

579: \setcounter{footnote}{0}

580: 

581: As in ref.\cite{Anchordoqui:2002cp} we randomly distribute the initial partons

582: inside a shell with thickness equal to the black hole lifetime $\tau_{bh}$

583: (\ref{life}) around the decayed black hole with radius \eqref{radius}. Quarks

584: and gluons are generated separately, with average multiplicities given by

585: Eq.(\ref{nav}) and energy distributions according to Fig.~1. The momenta of

586: most partons are chosen randomly inside that half of the solid angle which

587: points away from the black hole, as seen from the location at which the

588: particle is created. Choosing the partons to be always emitted radially by the

589: black hole, which would be proper for non--rotating black hole, would make

590: future collisions between them impossible in the absence of showering; our

591: choice therefore increases the possibility of parton--parton

592: scattering.\footnote{Non--radial emission should occur during the spin--down

593:   phase, with quite complicated angular dependence

594:   \cite{Ida1,kanti,Ida:2006tf}. Since most of the energy is released after the

595:   black hole has shed its spin, our treatment most likely over--estimates the

596:   probability of parton--parton scattering.} The 3--momentum of the last

597: parton is taken such that the total 3--momentum in the black hole rest frame

598: is zero.\footnote{This is not always possible, given that the energy of this

599:   parton has already been fixed and its virtuality must be smaller than this

600:   energy. If no solution is found, the event is abandoned, and the routine for

601:   generating the initial set--up makes a new attempt.} This is not absolutely

602: necessary, because the 3--momenta of the strongly interacting particles alone

603: do not have to add up to zero, but it allows a good control of the simulation.

604: 

605: Next, quark flavors are assigned randomly. We do not include the top and

606: bottom quark because we cannot assume them to be massless in all possible

607: collisions when we are using the massless QCD scattering amplitudes; indeed,

608: top production is likely to be somewhat suppressed since for our choice of

609: parameters, the Hawking temperature is somewhat below $m_t$. Charge

610: conservation is not enforced in our initial setup taking into account that the

611: black hole also radiates off other particles like leptons. However we take

612: care that the charges add up to some integer.

613: 

614: We set the maximal initial virtualities of the partons equal to their

615: energies; the actual values of these virtualities will be chosen by the shower

616: algorithm described in a subsequent Subsection. This choice of maximal

617: virtuality reproduces features of hadronic $Z$ decays fairly accurately, when

618: starting from leading order, $Z \rightarrow q \bar q$, decays.\footnote{One

619:   might argue that this choice is not appropriate for black hole decay, since

620:   black holes are not pointlike, unlike $Z$ bosons. However, for choices of

621:   $M_D$ and $M$ relevant for LHC phenomenology, the scales $1/r_h$ or

622:   $\Gamma_{bh}$ that describe the spatio--temporal extension of the black hole

623:   are quite close to the average energies of the decay partons; note also that

624:   the final results will only depend logarithmically on the maximal initial

625:   virtuality.}

626: 

627: 

628: After the initial setup the program enters the main loop, which simulates the

629: space--time evolution of the parton cascade. This evolution is determined by

630: two processes, parton scattering and branching, which are described in the

631: next two Subsections.

632: 

633: %----------------------------------------------------------------------

634: 

635: \subsection{Parton scattering}

636: \setcounter{footnote}{0}

637: 

638: Every possible pair of partons is boosted from the black hole rest frame into

639: its cms frame, and it is checked if it has reached its closest possible

640: distance. In the next step it is checked whether partons which reached their

641: closest distance undergo a collision. We are using the cascade approach

642: \cite{Geiger:1991nj} for this purpose, according to which a collision takes

643: place if the closest distance of the parton pair is within the radius defined

644: by the total cross section of the specific process,

645: %

646: \begin{equation} \label{collision}

647: |\mathbf r_{a} - \mathbf r_{b}|_{\rm min} \leq \sqrt{ \frac

648:  {\hat{\sigma}_{ab}} {\pi}} \, .

649: \end{equation}

650: %

651: 

652: If there is more than one possible scattering channel for two partons the

653: total cross section will be the sum of the cross sections for all possible

654: final states,

655: %

656: \begin{equation} \label{sumcross}

657: \hat{\sigma}_{ab} = \sum_{c,d} \hat{\sigma}_{ab \rightarrow cd} \, .

658: \end{equation}

659: %

660: The individual cross sections are calculated numerically by integrating the

661: corresponding differential cross sections

662: %

663: \begin{equation} \label{diffcross}

664: \hat{\sigma}_{ab \rightarrow cd} = \int_{\hat{t}_{\rm min}}^{\hat{t}_{\rm max}}

665: \left(\frac{d\hat{\sigma}(\hat{s}, \hat{t}, \hat{u})} {d\hat{t}} \right)_{ab

666:   \rightarrow cd} d\hat{t} \, ,

667: \end{equation}

668: %

669: with Mandelstam variables $\hat{s} = (p_a + p_b)^2$, $\hat{t} = (p_a -

670: p_c)^2$, $\hat{u} = (p_a - p_d)^2$. The choice of the upper and lower bounds

671: $\hat{t}_{\rm max}$, $\hat{t}_{\rm min}$ in equation \eqref{diffcross}

672: requires some care. When two on--shell partons are scattering, the matrix

673: elements for the ($2 \rightarrow 2$) QCD cross sections diverge for forward

674: ($\hat{t} \rightarrow 0$) and/or backward ($\hat{u} \rightarrow 0$)

675: scattering, thus a minimal momentum transfer is needed that determines

676: $\hat{t}_{\rm max}$ and $\hat{t}_{\rm min}$ for given $\hat{s}$. In this case

677: we take the commonly used value (in the parton--parton cms) of $p_{\perp {\rm

678:     min}} = 1$ GeV; this requires $\sqrt{\hat{s}} > 2$ GeV. Collisions with

679: $p_\perp < 1$ GeV are considered to be soft and are not evaluated since only

680: collisions generating high transverse momentum can change the jet structure of

681: the event.

682: 

683: Collisions with at least one virtual particle in the initial state need

684: special treatment, since off--shell initial or final partons lead to non gauge

685: invariant ($2 \rightarrow 2$) amplitudes \cite{Geiger:1991nj}. The authors of

686: \cite{Geiger:1991nj} solve this problem by combining the scattering with

687: space-- or time--like branching in a single (rather large) time step so that

688: at the end of a scattering only on--shell particles are left and only

689: on--shell particles will scatter again. This should be sufficient for the

690: simulation of heavy ion collisions, where (nearly) all virtualities and

691: exchanged 4--momenta are quite small. However, in our case this procedure

692: would allow a parton to instantaneously\footnote{Our program uses much shorter

693:   time steps than the original VNI code, typically $\sim 2 \cdot 10^{-4}$

694:   GeV$^{-1}$ initially. We checked that choosing even shorter steps does not

695:   change the result.} shower off a virtuality of hundreds of

696: GeV and then undergo a collision with momentum exchange of only a few GeV.

697: This does not make sense, since such a relatively soft collision would take

698: much more time than the (initial part of) the showering; moreover, the (many)

699: particles produced in the shower might undergo scatterings of their own.

700: 

701: We therefore take another approach: we allow virtual particles to take part in

702: a collision only if the scattering scale $Q^2_{\rm scatt}$ is at least as high

703: as half the sum of the virtualities $Q^2_a$, $Q^2_b$ of the particles $a$ and

704: $b$ in the initial state,

705: %

706: \begin{equation} \label{qscale}

707: Q^2_{\rm scatt} \geq \frac{Q^2_a + Q^2_b}{2} \, .

708: \end{equation}

709: %

710: In this case the scattering will at least not take longer than the parton

711: shower up to that point. Moreover, it can be hoped that the scattering is hard

712: enough that the virtuality of the particles in the initial state becomes

713: irrelevant, so that we can describe the scattering using massless ($2

714: \rightarrow 2$) QCD cross sections. In fact, basically the same condition is

715: chosen by the usual QCD simulators when setting the showering scale for

716: initial state radiation, although in these programs the scattering scale

717: $Q^2_{\rm scatt}$ is fixed first. In our simulation the scattering scale is

718: taken to be

719: %

720: \begin{equation} \label{qscatt}

721: Q^2_{\rm scatt} = \frac {\hat{t} \hat{u}} {\hat{s}} 

722: \simeq \mathbf p_{\perp}^2 \, .

723: \end{equation}

724: %

725: This also fixes the scale in the running strong coupling constant,

726: %

727: \begin{equation} \label{alphas}

728: \alpha_s(Q^2_{\rm scatt}) = \frac{12\pi}{25\ \log(\frac{Q_{\rm scatt}^2}

729:   {\Lambda_{QCD}^2})} \, ,

730: \end{equation}

731: %

732: where we took $N_F = 4$ active flavors and QCD scale $\Lambda_{QCD} = 0.2$ GeV.

733: The requirement \eqref{qscale} determines the boundaries of the

734: integration over $\hat{t}$ in \eqref{diffcross}:

735: %

736: \begin{equation} \label{tbounds}

737: \hat{t}_{\rm min, \, max} = \frac{1}{2} \left( -\hat{s} \mp \sqrt{\hat{s}^2 - 

738:     4Q^2_{ab}\hat{s}} \right) \, ,

739: \end{equation}

740: % 

741: where $Q^2_{ab} = \frac{Q^2_a + Q^2_b}{2}$. 

742: 

743: We did not include ($2 \rightarrow 1$) fusion processes like $g + g

744: \rightarrow g^*$ (with $g^*$ being off--shell), since the first branching of

745: the produced off--shell parton would again result in a $2 \rightarrow 2$

746: process, leading to double counting. It could be argued that we should also

747: require $Q_{\rm scatt} \geq 1 / |\mathbf r_{a} - \mathbf r_{b}|_{\rm min}$.

748: However, this would exclude interactions between partons that happen somewhat

749: before or after the time of their closest approach, which seems unphysical. We

750: therefore do not impose this additional requirement, which would reduce the

751: number of partonic scatterings significantly. In fact, the opposite

752: requirement, $Q_{\rm scatt} < 1 / |\mathbf r_{a} - \mathbf r_{b}|_{\rm min}$,

753: seems more reasonable, since according to the uncertainty principle a highly

754: virtual exchanged parton can only travel a very short distance. We do not

755: impose this constraint, i.e. we allow very large momentum exchange also

756: between relatively distant partons. This again increases the importance of

757: collisions, since a large $Q_{\rm scatt}$ is required for a collision to

758: modify the jet structure. However, we will see that not imposing this upper

759: bound on $Q_{\rm scatt}$ has very little effect in practice, since high values

760: of $Q_{\rm scatt}$ are in any case very unlikely.

761: 

762: In some rare cases one parton has two possible collision partners in a single

763: time step. In that case for both partners cross section and distance will be

764: checked. If only one collision takes place this will be the one evaluated. If

765: both collisions would take place only the one with the larger cross section

766: will be evaluated.

767: 

768: After a pair of particles has been identified as colliding, the scattering

769: kinematics can be generated. For initial states which have more than one

770: channel, the final state is chosen according to the probability given by the

771: relevant total cross sections. Next the value of $\hat{t} \in \left[

772:   \hat{t}_{\rm min}, \, \hat{t}_{\rm max} \right]$ is generated, with

773: distribution given by the corresponding differential cross section $d \hat

774: \sigma/ d \hat t$ (normalized to unity). The initial particle pair is boosted

775: into its cms frame and put on--shell. The absolute value of the transverse

776: momentum of the final particles is given by $Q^2_{\rm scatt}$ as determined by

777: the chosen value of $\hat{t}$. The azimuthal scattering angle, which fixes the

778: direction of the $\mathbf{p_\perp}$ vectors, is chosen randomly, with flat

779: distribution between 0 and 2$\pi$.

780: 

781: Finally the virtualities of the outgoing particles are created by the

782: branching routine (see the next Subsection) and color charges are assigned

783: according to the color flow of the specific process. It should be noted that

784: in case of the scattering of virtual particles, the virtuality after

785: scattering can become even lower than before, since $Q_{\rm scatt}$ is the

786: {\em maximal} virtuality of the particles in the final state. We see no

787: physical reason why such virtuality--reducing scattering reactions should be

788: suppressed. 

789: 

790: Due to the lifetime of a virtual (exchanged) particle the scattering will take

791: a finite amount of time determined by the scattering scale. In the rest frame

792: of the black hole this time is given by

793: %

794: \begin{equation} \label{tscatt}

795: \tau_{\rm scatt} = \frac{\gamma_{ab}}{Q_{\rm scatt}} \, ,

796: \end{equation}

797: %

798: where $\gamma_{ab} = (E_a + E_b) / \sqrt{\hat{s}_{ab}}$ is the boost factor

799: between the cms of colliding partons $a$ and $b$ and the black hole rest

800: frame. During this time the colliding particles are not allowed to collide

801: again, nor can they branch off additional partons; in fact, in quantum

802: mechanics one cannot say whether the parent partons $a,b$ or the children

803: $c,d$ exist during the period $\tau_{\rm scatt}$. After the scattering time is

804: over the partons are able to initiate final state radiation or collide

805: again.\footnote{In principle it might be more appropriate to place this ``dead

806:   time'' symmetrically around the time of closest approach, rather than

807:   letting it start at the time of closest approach. However, this would be

808:   technically difficult, since the program would then have to go back in time,

809:   and ``un--do'' any possible branchings that happened in the ``dead'' period

810:   before the time of closest approach. Since by construction $Q_{\rm scatt}$

811:   is larger than the virtuality of the scattering particles, this asymmetric

812:   placement of the ``dead time'' should not matter much in practice.}

813: 

814: %-----------------------------------------------------------------------

815: 

816: \subsection{Parton branching}

817: \setcounter{footnote}{0}

818: 

819: While the inclusion of partonic collisions is the main new ingredient of our

820: simulation, we will see that parton branching plays a far more important role

821: in determining the characteristics of the final state. We model this using a

822: modified version of the relevant routine of VNI 4.12b, which in turn is based

823: on the PYTHIA branching algorithm \cite{Sjostrand:2003wg, Bengtsson:1986et,

824:   Webber:1986mc}.

825: 

826: The modifications implemented by us address the need to do branching for a

827: single parton, instead of treating two partons at once as is done normally in

828: order to ensure 4--momentum conservation. To make sure we can conserve energy

829: and momentum for a single parton we always have to determine the virtualities

830: at which the parton branches one step ahead in the branching algorithm: the

831: kinematics of $a \rightarrow b + c$ can only be fixed if the ``masses'', i.e.

832: (time--like) virtualities, of all three participating partons are known. Our

833: version of the routine therefore determines the virtualities of $b$ and $c$

834: earlier than the original routine does. Note that these actual virtualities

835: are typically much smaller than the corresponding maximal values; this is the

836: reason why energetic partons, with initial {\em maximal} virtualities given by

837: their energies, typically produce quite narrow jets. Another modification is

838: the introduction of the scattering--induced ``dead time'' described at the end

839: of the previous Subsection, during which partons are not allowed to branch.

840: 

841: Just like scattering reactions, the branching $a \rightarrow b + c$ also takes

842: a finite amount of time in our simulation, given by

843: %

844: \begin{equation} \label{tbranch}

845: \tau_{\rm branch} = \frac{\gamma_a}{Q_a} = \frac{E_a}{Q_a^2} \, ,

846: \end{equation}

847: %

848: with $\gamma_a = E_a/Q_a$ describing the boost from the rest frame of parton

849: $a$ to that of the black hole. In our simulation this is treated like the

850: lifetime of a decaying particle, i.e. the probability of an ``active'' parton

851: $a$ to branch during the time step $dt$ is given by \cite{Geiger:1998fq,

852:   Bass:1999pv}

853: %

854: \begin{equation} \label{pbranch}

855: P_{\rm branch} = 1 - {\rm e}^{-dt/\tau_{\rm  branch}} \simeq \frac {dt}

856:     {\tau_{\rm branch}} \, ,

857: \end{equation}

858: %

859: where the second, approximate equality holds for the (realistic) situation $dt

860: \ll \tau_{\rm branch}$. Note that each parton is checked for possible

861: scattering before it is checked for branching. In our treatment partons $b,c$

862: are created instantaneously if a branching occurs, i.e. they are allowed to

863: undergo scattering immediately.\footnote{These partons are created at slightly

864:   different locations, as determined by the uncertainty principle; moreover,

865:   they are moving away from each other. These two partons can therefore not

866:   scatter on each other, even if they are very close, but they can scatter on

867:   other partons.} This can be justified from the requirement (\ref{qscale}),

868: which ensures that scattering reactions are much faster than branchings.

869: Imposing a branching ``dead time'' of (some fraction of) $\tau_{\rm branch}$

870: on $b,c$ before they are allowed to scatter would obviously reduce the number

871: of parton--parton scatterings, although numerically this effect is not very

872: large. The 4--momenta of partons $b,c$ are chosen as in the PYTHIA branching

873: algorithm \cite{Sjostrand:2003wg}.

874: 

875: There are three situations in which the branching routine is used. First, all

876: initial partons which have some maximum virtuality from black hole decay are

877: sent through the branching routine once, in order to determine their actual

878: starting virtuality for the simulation. Secondly, the normal use when every

879: time step all partons which have any time--like virtuality left, and which did

880: not scatter recently, are sent through the routine to undergo branching with

881: probability given by Eq.(\ref{pbranch}). Third, similar to the first case, the

882: two partons coming out of a hard scattering which took place at the scale

883: $Q^2_{\rm scatt}$ are sent through the branching routine to determine the

884: actual virtualities they will start to propagate with.

885: 

886: The next step after the parton branching algorithm is to propagate all partons

887: freely by one time step. Then the simulation calculates the average distance

888: between all pairs of partons to see if it exceeds the QCD scale

889: $1/\Lambda_{QCD}$, at which point the partonic simulation loop can be aborted.

890: If the abort condition is met hadronization will take over, or else time is

891: increased by one time step and the simulation loop starts again. For details of

892: the hadronization algorithm we again refer to the documentation of PYTHIA

893: \cite{Sjostrand:2003wg}.

894: 

895: %------------------------------------------------------------------------

896: 

897: \section{Results}

898: 

899: We are now ready to present some numerical results. As stated in Sec.~2, we

900: set the higher--dimensional Planck scale to its lower bound, $M_D = 0.65$ TeV,

901: since this maximizes the number of initial partons, and hence the number of

902: parton--parton interactions within a given black hole decay. We also consider

903: relatively heavy black holes, with $M = 5$ and 10 TeV. Recall that decays of

904: heavier black holes are characterized by lower temperatures, and hence higher

905: multiplicities; however, the production cross section for black holes with

906: mass exceeding 10 TeV is certainly negligible at the LHC.

907: 

908: Since the initial set--up is created totally randomly there are large

909: event--to--event fluctuations of the number and energies of the initial

910: partons. One therefore needs sufficient statistics to make reliable statements

911: about average quantities; the results presented below are based on 200 events

912: for each run. In order to illustrate the effects of parton--parton

913: interactions, we made separate runs for each black hole mass where these

914: interactions were turned on or turned off. In the latter case the

915: spatio--temporal evolution of the parton shower is irrelevant, i.e. our

916: program reproduces standard (PYTHIA) showering for the given set--up of

917: original partons.

918: 

919: If black hole decays led to formation of a chromosphere, observables based on

920: jets would no longer be useful. In the following Subsections we therefore

921: analyze the distribution of two observables whose definition does not assume

922: the existence of jets: angular correlations between energetic charged hadrons,

923: and the overall energy flow. Of course, the observables are chosen such that

924: their distribution would be greatly affected if parton--parton scattering did

925: indeed lead to formation of a chromosphere. In the last Subsection we will

926: interpret these results with the help of the time structure of the parton

927: shower that develops after typical black hole decays.

928: 

929: %---------------------------------------------------------------------

930: 

931: \subsection{Angular correlation}

932: \setcounter{footnote}{0}

933: 

934: After hadronization we are left with a large number of charged, long--lived

935: particles (mostly charged pions and kaons and some protons) as well as photons

936: and neutral hadrons. In this first analysis we focus on charged particles

937: because their momenta can be measured accurately using tracking information.

938: We only consider particles with an energy (in the black hole rest frame) above

939: 4 GeV; the number of charged particles passing this cut is denoted by $N_{\rm

940:   ch}$. This cut should largely remove particles from the underlying event

941: which have nothing to do with black hole decay, and which have not been

942: included in our simulation.

943: 

944: We compute the angles $\theta$ between any two of these charged particles;

945: altogether there are $N_{\rm pair} = N_{\rm ch}(N_{\rm ch} - 1) / 2$ such

946: pairs. The resulting distribution is binned in $\cos\theta$. Because $N_{\rm

947:   ch}$ can vary a lot from event to event, we normalize the distribution for

948: each event to $N_{\rm pair}$, before averaging over the 200 generated events.

949: 

950: Let us first consider some simple situations, in order to get a feeling for

951: what kind of distribution to expect. The simplest case, which certainly does

952: not describe black hole decay, would be events with two back--to--back jets,

953: each of which contains the same number $N_{\rm ch}/2$ of charged particles. In

954: this case $N^2_{\rm ch} / 4 - N_{\rm ch} / 2$ pairs would have an angle near

955: zero between the particles of the pair because they reside inside the same

956: jet; the remaining $N^2_{\rm ch} / 4$ pairs would have an angle near $\pi$

957: between the particles of the pair because the particles are in different jets.

958: In the limit of large $N_{\rm ch}$ these numbers will become almost equal,

959: leading to peaks at $1$ and $-1$ in our plot which have approximately the same

960: height.

961: 

962: A case that is closer to what one may expect from the decay of a black hole

963: would be an event with $n_j$ jets. Let us keep the assumption that each of

964: these jets contains exactly the same number $N_{\rm ch}/n_j$ of energetic

965: charged particles. In this case there are

966: %

967: \begin{equation} \label{pospairs}

968: \frac{N_{\rm ch}} {2} \left(\frac{N_{\rm ch}}{n_j} - 1 \right)

969: \end{equation}

970: %

971: pairs residing in the same jet; the remaining $N^2_{\rm ch} (n_j -

972:   1) / (2n_j)$ pairs of charged particles are in different jets. We thus see

973: that the fraction of all pairs that reside inside the same jet will decrease

974: when the number of jets increases, i.e. the peak in our distribution at

975: $\cos\theta = +1$ will become smaller. Moreover, for $n_j \gg 2$ randomly

976: distributed jets the peak at $\cos\theta = -1$ will vanish because momentum

977: conservation no longer requires any two jets in the event to be

978: back--to--back. After averaging over many events all the angles between

979: different jets contribute more or less equally. In this case we therefore

980: expect a smooth distribution (although momentum conservation my still lead to

981: a slight increase of the correlation function at negative $\cos\theta$) with

982: only one peak at $\cos\theta = +1$. The distribution should remain

983: qualitatively the same in the more realistic scenario where we allow different

984: jets to contain different numbers of charged particles, although

985: eq.(\ref{pospairs}) will then no longer be valid.

986: 

987: On the other hand, if we assume that a chromosphere is indeed spherical, the

988: correlation function should become almost perfectly flat; in particular, we

989: would not expect any visible peak at $\cos\theta = +1$.

990: 

991: Figs.~2 show results for black hole mass $M = 5$ TeV (top) and 10 TeV

992: (bottom), with (red) and without (black) parton--parton scattering.  Clearly

993: for both black hole masses there is still a strong peak at $\cos\theta = 1$,

994: leading to the conclusion that we should expect a jet structure after black

995: hole decay. The peak becomes smaller for a heavier black hole; this is

996: expected from the qualitative discussion presented above, given the increasing

997: initial parton multiplicity, see Eq.(\ref{nav}). The rise of the distribution

998: towards $\cos\theta=-1$ is also more pronounced for smaller black hole masses,

999: again as expected from our qualitative discussion. For $M=5$ TeV,

1000: parton--parton scattering has essentially no effect on this distribution. For

1001: $M=10$ TeV, it leads to a very slight broadening of the peak at

1002: $\cos\theta=+1$; however, given the uncertainties of our simulation, we do not

1003: claim that this effect is significant.

1004: 

1005: \begin{figure}[H]

1006: \vspace*{-2cm}

1007: \begin{center}

1008: \includegraphics[scale=1.3]{5tev-1-new3.eps}\\

1009: \vspace*{-1.5cm}

1010: \includegraphics[scale=1.3]{10tev-1-new3.eps}

1011: \end{center}

1012: \vspace*{-5mm}

1013: \caption{The angular correlation function for charged particles from black

1014:   hole decay with $E>4$ GeV, for black hole mass $M = 5$ (top) and 10 TeV

1015:   (bottom), obtained by binning into 1,000 bins. The red and black curves have

1016:   been obtained including and omitting parton--parton scattering,

1017:   respectively.}

1018: \end{figure} \label{fig_angle}

1019: 

1020: %-----------------------------------------------------------------------

1021: 

1022: \subsection{Total energy flow}

1023: \setcounter{footnote}{0}

1024: 

1025: The second quantity we investigate is the total energy flow from the hadronic

1026: black hole decay products. Here we envision a calorimetric measurement. We

1027: therefore divide phase space into azimuthal angle $\phi \in [0, 2\pi]$ and

1028: pseudo--rapidity $\eta \in [-4,4]$, with 15 bins in $\phi$ and 30 bins in

1029: $\eta$. For each of these 450 ``calorimeter cells'' the total visible energy

1030: is calculated, including hadrons and photons from hadronic decay, but no

1031: neutrinos or muons. We present the result by plotting the number of cells with

1032: deposited energy $E_{\rm cell} \leq E_{\rm max}$ as function of $E_{\rm max}$.

1033: 

1034: Let us again first discuss the possible shapes of this distribution for

1035: different final parton configurations. If a chromosphere forms, we expect a

1036: very large number of hadrons in the final state, each of which has a

1037: relatively small energy. These would be distributed uniformly over phase

1038: space, i.e. had we defined our cells as having constant length in $\cos\theta$

1039: all cells would receive essentially the same energy. We prefer to use $\eta$

1040: to parameterize the phase space, since the energy flow pattern will then be

1041: invariant under motion of the black hole along the beam pipe. Since $d / d

1042: \eta = \sin^2(\theta) d / d \cos\theta$, cells at small $\cos\theta$, i.e.

1043: small $|\eta|$, will then receive significantly more energy than those at

1044: large $|\eta|$. There should nevertheless be almost no empty cells; the

1045: maximal energy deposit, in cells with $\eta \sim 0$, would be about $0.009

1046: E_{\rm tot}$, where $E_{\rm tot} \simeq 0.75 M$ is the total hadronic energy

1047: released in the decay of a black hole with mass M. In particular, there should

1048: not be any cells with energy comparable to the initial average partonic

1049: energy, which amounts to about 300 GeV for our choices of parameters [see the

1050: discussion of Eq.(\ref{nav}) in Sec.~2]. The distribution expected for

1051: ``ideal'' chromospheres, with energy flow being completely independent of

1052: $\phi$ and $\cos\theta$, is depicted by the blue step--like\footnote{These

1053:   steps appear only for black hole decays at rest. In general one expects a

1054:   smoothed-out version of these curves once a distribution of longitudinal

1055:   momenta of the black holes is taken into account.} curves in Figs.~3.

1056: 

1057: In the opposite extreme, where the final state consists of a relatively small

1058: number [given by Eq.(\ref{nav})] of very narrow jets, most cells would be

1059: empty, while in a few cells the deposited energy would be of order 300 GeV.

1060: However, even in the absence of parton--parton scattering, final state

1061: radiation implies that many jets will spread out over several cells. Together

1062: with the final hadronization step, this will lead to a significant number of

1063: cells in which a small, but nonzero, amount of energy is deposited, often in

1064: form of a single hadron. We therefore expect a non--trivial dependence on

1065: $E_{\rm max}$ in the entire range between $\sim 100$ MeV and 1 TeV.

1066: 

1067: Our results for the energy flow of hadronic black hole decay products are

1068: shown in Figs.~3. We see that the number of cells with $E < E_{\rm max}$

1069: indeed shows nontrivial dependence on $E_{\rm max}$ over a wide range. The

1070: number of (almost) empty cells decreases with increasing black hole mass, as

1071: expected from the higher initial parton multiplicity (\ref{nav}).  However,

1072: even for $M=10$ TeV, in about 50\% of the cells almost no energy is deposited;

1073: on the other hand, about twenty cells contain more than 100 GeV. We saw above

1074: that these results are consistent with the existence of well--defined jets.

1075: Parton--parton scattering has practically no effect on the number of cells

1076: containing at least 30 GeV, again indicating that it does not affect the jet

1077: structure at all.

1078: 

1079: 

1080: \begin{figure}[H]

1081: \begin{center}

1082: \vspace*{-2cm}

1083: \includegraphics[scale=1.3]{5tev-2-new3.eps}\\

1084: \vspace*{-1.5cm}

1085: \includegraphics[scale=1.3]{10tev-2-new3.eps}

1086: \end{center}

1087: \vspace*{-5mm}

1088: \caption{The number of phase space cells where the deposited energy satisfies

1089:   $E < E_{\rm max}$ as function of $E_{\rm max}$, for black hole mass $M = 5$

1090:   TeV (top) and 10 TeV (bottom). The notation is as in Fig.~2, except that the

1091:   blue, step--like, curves show the prediction from an ideal chromosphere.}

1092: \end{figure} \label{fig_energy}

1093: 

1094: Scattering does increase the total number of non--empty cells by about 5\%

1095: (10\%) for $M=5$ (10) TeV. However, the production of soft hadrons is a

1096: non--perturbative process, and can therefore not be treated from first

1097: principles; it is not clear whether the effect of parton--parton scattering is

1098: larger than the systematic error of our simulation.\footnote{The {\em

1099:     relative} size of the effect of parton scattering should be rather

1100:   insensitive to the details of the simulation; the statement that it

1101:   increases the number of non--empty cells by 5 to 10\% should therefore be

1102:   relatively robust. However, at the end one can only compare the {\em

1103:     absolute} prediction of the simulation with actual events.} Moreover, it

1104: should be kept in mind that the ``underlying event'', which is created by the

1105: remnants of the colliding protons that do not participate in black hole

1106: formation, will also contribute a large number of (mostly soft) hadrons to the

1107: final state, which have not been included in our simulation. It is therefore

1108: not clear whether the increase of the cells containing some black hole decay

1109: products results in a measurable difference in the total energy flow in the

1110: event.

1111: 

1112: %----------------------------------------------------------------------

1113: 

1114: \subsection{Microscopic structure of the events}

1115: \setcounter{footnote}{0}

1116: 

1117: The results of the two previous Subsections show that parton--parton

1118: interactions have little effect on the global characteristics of the hadronic

1119: final state that results from the decay of black holes with masses of a few

1120: TeV. In this Subsection we analyze the microscopic structure of the evolution

1121: of the hadronic final state that results from the decay of such black holes.

1122: 

1123: In Table~1 we list the average parton multiplicities (just before

1124: hadronization), as well as the average number of parton--parton collisions,

1125: for $M=5$ and 10 TeV. We see that even in the absence of parton--parton

1126: scattering, QCD branching (final state radiation) increases the multiplicity

1127: by an order of magnitude, relative to the initial multiplicity given by

1128: Eq.(\ref{nav}). We also see that the average number of parton--parton

1129: scatterings per black hole decay is not so small; as expected from simple

1130: statistical arguments, it increases roughly quadratically with the (initial or

1131: final) partonic multiplicity. Scattering increases the final parton

1132: multiplicity by 16\% (23\%) for $M=5$ (10) TeV. This effect is even larger

1133: than that on the number of non--empty calorimeter cells. Note, however, that

1134: the number of scatterings still remains well below the number of partons even

1135: for $M=10$ TeV.

1136: 

1137: \begin{table}[h]

1138: \begin{center}

1139: \begin{tabular}{|c|| c | c | c || c| c| c| c|}

1140: \hline

1141: $M$ [TeV] & \multicolumn{3}{c||}{w/o scattering } & 

1142: \multicolumn{4}{c|}{with scattering } \\

1143: & $\langle n_q \rangle$ &  $\langle n_g \rangle$ & $\langle n_{\rm parton}

1144: \rangle$ &$\langle n_q \rangle$ &  $\langle n_g \rangle$ & $\langle n_{\rm

1145:   parton} \rangle$ & $\langle n_{\rm scatt} \rangle$ \\

1146: \hline

1147: 5  & 19 & 116 & 135 & 21 & 134 & 156 & 16 \\

1148: 10 & 41 & 241 & 282 & 50 & 295 & 346 & 53 \\

1149: \hline

1150: \end{tabular}

1151: \end{center}

1152: \caption{Average final quark, gluon and total parton multiplicities from the

1153:   decay of black holes with 5 and 10 TeV mass, with and without including the

1154:   effect of parton--parton scattering. The last column gives the average

1155:   number of partonic collisions. The statistical errors on the average

1156:   multiplicities are about 2 percent.}

1157: \end{table}

1158: 

1159: On the other hand, the probability $\langle n_{\rm parton} \rangle / \langle

1160: n_{\rm scatt} \rangle$ that a given parton resulted from scattering is not

1161: negligible. It rises roughly linearly with the parton multiplicity, in

1162: agreement with the estimate (\ref{intrate}). However, this expression, with

1163: $Q_{\rm min} \sim 1$ GeV as in our simulation, greatly over--estimates the

1164: number of scattering reactions even if we normalize it to the initial parton

1165: multiplicity.

1166: 

1167: As mentioned earlier, parton--parton scattering can only destroy the jet

1168: structure if it involves large momentum exchange, i.e. if the scale $Q_{\rm

1169:   scatt}$ defined in Eq.(\ref{qscatt}) is large. In Fig.~4 we show a scatter

1170: plot of $Q_{\rm scatt}$ values vs. time in the simulation; here entries from

1171: 100 decays of black holes with mass $M=5$ TeV have been collected. We see

1172: that the average value of $Q_{\rm scatt}$ decreases with time, once $t > 1/

1173: \langle E_{\rm parton} \rangle \sim 3 \cdot 10^{-2}$ GeV$^{-1}$. This can be

1174: understood from the observation that at early times, most partons are quite

1175: far off--shell; our condition (\ref{qscale}) then implies that early

1176: scatterings must involve rather large momentum exchange. The existence of a

1177: few early scatterings with low $Q_{\rm scatt}$ is due to the fact that the

1178: initial virtualities with which the partons are actually created by the

1179: simulation may be much smaller than their maximal values, which is given by

1180: the energies of these partons.

1181: 

1182: \begin{figure}[H]

1183: \begin{center}

1184: \includegraphics[scale=1.2]{scale.eps}

1185: \end{center}

1186: \caption{Scatter plot showing the scale $Q_{\rm scale}$ defined in

1187:   Eq.(\ref{qscatt}) vs. time after black hole decay, for $M=10$ TeV;

1188:   scatterings from 100 decays have been added into this plot. The appearance

1189:   of discrete lines at early times is an artefact of the finite time step size

1190:   used.}

1191: \end{figure} \label{fig_qscale}

1192: 

1193: Since the average virtualities of the partons diminish with increasing time,

1194: more scatterings with small $Q_{\rm scatt}$ become possible. Since the QCD

1195: cross sections satisfy $d \sigma / d Q^2_{\rm scatt} \propto \alpha_s(Q_{\rm

1196:   scatt})^2 Q^{-4}_{\rm scatt}$ if $\hat t \hat u \ll \hat s^2$, reactions

1197: with large $Q_{\rm scatt}$ are then greatly disfavored. Moreover, the average

1198: parton energies also decrease with time. Later collisions therefore tend to

1199: have smaller $\hat{s}$, which implies reduced kinematical upper bounds on

1200: $Q_{\rm scatt}$.

1201: 

1202: The time dependence of the number of scattering reactions is determined by

1203: several competing effects. While the number of partons increases due to

1204: multiple branchings as time goes on, the number {\em density} of partons

1205: decreases, making it increasingly less likely that two partons will come close

1206: to each other. As a result, the number of collisions per unit time decreases

1207: at first. On the other hand, the decreasing virtuality of the partons,

1208: and corresponding decreasing lower bound on $Q_{\rm scatt}$, means that their

1209: scattering cross section increases with time. This leads to an accumulation of

1210: relatively soft scatterings at $t \sim 0.1/$GeV.

1211: 

1212: A typical decay of a 5 TeV black hole therefore develops as follows.

1213: Initially there are about 12 highly off--shell partons with average energies

1214: around 300 GeV. One or two of them may undergo scattering very soon after

1215: their creation, with typical scattering scale $Q_{\rm scatt}$ of a few tens of

1216: GeV.  Note that these rather hard reactions still mostly have $Q_{\rm scatt}

1217: \ll E$, so that the direction of the participating partons is not changed very

1218: much.  Moreover, these early scatterings may actually reduce the virtualities

1219: of the participating partons. As time goes on, the number of partons in the

1220: event increases, mostly due to branching processes. The number of scatterings

1221: also increases, but most of these reactions are relatively soft. We saw in

1222: Table~1 that they do increase the partonic multiplicity at the end of the QCD

1223: cascade.  However, most of these scattering reactions involve partons inside

1224: the same jet, and therefore have little effect on the global quantities we

1225: analyzed in the previous Subsections. In fact, the creation of an additional

1226: very soft parton nearby in phase--space may not be visible in the (hadronic)

1227: final state at all. This explains why the results in Table~1 seem to indicate

1228: larger effects from parton--parton scattering than what we saw in Figs.~2 and

1229: 3.

1230: 

1231: \section{Summary and conclusions}

1232: 

1233: In this paper we investigated the question if the decays of black holes that

1234: might be produced at the LHC would be characterized by a relatively large

1235: number of discrete jets, as is usually assumed, or if they would lead to the

1236: formation of quasi--thermal ``chromospheres'' of rather soft hadrons, as

1237: suggested in ref.\cite{Anchordoqui:2002cp}. To this end we first described the

1238: partonic state produced by the decay of a black hole. We found that, if the

1239: higher--dimensional Planck mass $M_D$ is at its current lower bound, the

1240: average initial number of partons will be less than about 25 for black holes

1241: with mass $M \leq 10$ TeV, which might have a chance to be produced with

1242: appreciable rates at the LHC; higher values of $M_D$ would lead to initial

1243: states with fewer partons, and hence to less scattering.

1244: 

1245: In Sec.~3 we summarized the argument in favor of formation of a

1246: chromosphere. We pointed out that this argument ignores the fact that

1247: scattering reactions take a finite amount of time. Moreover, a very similar

1248: argument should apply to ordinary QCD multi--jet events. It would predict

1249: sizable effects in six jet events, which have been studied experimentally by

1250: the UA2 and CDF collaborations, who found good agreement with standard QCD

1251: predictions which ignore interactions between the partons in the final state.

1252: 

1253: Clearly a dedicated Monte Carlo study is required in order to determine the

1254: quantitative effects of such final state interactions on black hole events at

1255: the LHC. As described in Sec.~4, we wrote such a simulation code, based on the

1256: VNI program \cite{Geiger:1998fq, Bass:1999pv}. We modified it by forcing

1257: scattering reactions to take a finite amount of time, which we estimated using

1258: the uncertainty principle. A very similar argument implies that, as time goes

1259: on, the parton system evolves towards smaller virtualities through branching

1260: processes. Scattering reactions may increase the virtualities of the

1261: participating partons again, but can also be a shortcut towards smaller

1262: virtualities. 

1263: 

1264: Results of our simulation are described in Sec.~5. We found the effects of

1265: parton--parton scattering in the final state to be essentially negligible both

1266: for the angular correlation between energetic charged hadrons, and for the

1267: number of phase space cells containing a large amount of energy. We found some

1268: effect on the number of cells containing only one or a few soft hadrons. We

1269: interpreted these results in terms of the microscopic structure of the event.

1270: In particular, we saw that ``hard'' reactions, with large momentum exchange,

1271: only occur early on; in this case they may well reduce the virtualities of the

1272: participating partons. Later scatterings are all quite soft. At both early and

1273: late times, most scattering reactions have momentum exchange well below the

1274: energies of the participating partons; such reactions cannot significantly

1275: change the directions into which these partons are traveling. We therefore

1276: conclude that a chromosphere will {\em not} form; programs that ignore

1277: interactions between the partons from the decay of black holes created at the

1278: LHC only make a negligible mistake.

1279: 

1280: Our simulation has some shortcomings. Like all QCD simulation codes, it

1281: essentially works on the level of squared matrix elements, i.e. quantum

1282: mechanical interference effects can only be included in an approximate

1283: manner. In addition, we had to identify the quantum mechanical uncertainty in

1284: time with actual duration, both for the branching and scattering processes. It

1285: is not clear to us how this limitation can be overcome, even in principle.

1286: 

1287: Our results starkly contradict the claims of ref.\cite{Anchordoqui:2002cp},

1288: even though we chose our parameters and initial set--up such as to maximize

1289: the likelihood of scattering reactions between partons in the final state. It

1290: might therefore be worthwhile to summarize the three effects that have reduced

1291: the number of such reactions in our simulation.

1292: 

1293: Perhaps most important is that we only allow scattering of partons that

1294: approach each other. Interactions between partons that ``always'' (after their

1295: creation) move away from each other can certainly not be treated using the

1296: factorization into initial black hole decay and subsequent parton--parton

1297: scattering that underlies our simulation; for one thing, no $S-$matrix could

1298: be defined for such a state.\footnote{In some sense such partons can

1299:   nevertheless still interact with each other. For example, the well--known

1300:   Coulomb singularity \cite{schwinger} for near--threshold decays into charged

1301:   particles can be interpreted in terms of the electromagnetic interaction

1302:   between these particles. However, in the simplest case of two--body decays,

1303:   this singularity only affects the partial width for the decay; it does not

1304:   increase the final state multiplicity. In case of decays into more than two

1305:   partons, it can also change kinematical distributions, e.g. the invariant

1306:   mass distributions of subsets of partons. However, a proper treatment of

1307:   these effects would require the inclusion of corrections to black hole

1308:   decays due to gauge loops. It is currently not clear how this could be

1309:   done.}

1310: 

1311: Strictly speaking, in our case one cannot define an $S-$matrix for partons

1312: approaching each other, either, since they were created at a finite time, i.e.

1313: they did not exist in an ``in-''state defined at time $t=-\infty$. We

1314: circumvent this problem by requiring the 4--momentum exchanged in these

1315: reactions to be larger than the initial virtualities. In this case one should

1316: be able to approximately treat the scattering like that of on--shell

1317: particles. This roughly means that we require scattering reactions to be fast

1318: on the time scale of the shower evolution up to the time of the scattering.

1319: This constraint greatly reduces the scattering cross section at early times,

1320: when the parton density is highest, and therefore reduces the number of

1321: scattering reactions. The fact that we do not allow partons participating in a

1322: scattering to start another scattering while the first one is still ``in

1323: progress'' has a much smaller effect on the number of these reactions.

1324: 

1325: Although the number of parton--parton collisions remains small for all black

1326: holes that might be produced at the LHC, we saw that it increases roughly

1327: quadratically with increasing black hole mass. Extrapolating from the results

1328: of Table~1, we estimate that there will be ${\cal O}(1)$ parton--parton

1329: scatterings per produced parton once $M \gsim 25$ TeV, if the

1330: higher--dimensional Planck scale $M_D$ is kept at its lower bound of 0.65

1331: TeV. This need not lead to formation of a chromosphere, however; we saw that

1332: most scattering reactions have little effect on the jet structure of the event.

1333: Note that in this case the average initial partonic multiplicity already

1334: exceeds 100, making the reconstruction of distinct jets quite unlikely, even

1335: if such heavy black holes could ever be produced at human--made colliders.

1336: 

1337: \subsubsection*{Acknowledgments}

1338: 

1339: The work of K.O. is partly supported by the Special Postdoctoral Researchers

1340: Program at RIKEN.

1341: 

1342: \begin{thebibliography}{99}

1343: 

1344: \bibitem{Arkani-Hamed:1998rs}

1345:   N.~Arkani-Hamed, S.~Dimopoulos and G.~R.~Dvali,

1346: %  ``The hierarchy problem and new dimensions at a millimeter,''

1347:   Phys.\ Lett.\ B {\bf 429} (1998) 263, hep--ph/9803315;

1348: I. Antonaidis, N. Arkani-Hamed, S. Dimopoulos and G.R. Dvali, Phys. Lett. B

1349: {\bf 436} (1998) 257, hep--ph/9804398.

1350: 

1351: \bibitem{Randall:1999ee}

1352:   L.~Randall and R.~Sundrum,

1353: %  ``A large mass hierarchy from a small extra dimension,''

1354:   Phys.\ Rev.\ Lett.\  {\bf 83} (1999) 3370, hep--ph/9905221.

1355: 

1356: \bibitem{'tHooft:1987rb}

1357:   G.~'t Hooft,

1358: %  ``Graviton Dominance In Ultrahigh-Energy Scattering,''

1359:   Phys.\ Lett.\ B {\bf 198} (1987) 61.

1360: 

1361: \bibitem{banks}

1362: T. Banks and W. Fischler, hep--th/9906038.

1363: 

1364: \bibitem{Giddings:2001bu}

1365:   S.~B.~Giddings and S.~D.~Thomas,

1366: %  ``High energy colliders as black hole factories: The end of short distance

1367: %  physics,''

1368:   Phys.\ Rev.\ D {\bf 65} (2002) 056010, hep--ph/0106219.

1369: 

1370: \bibitem{Dimopoulos:2001hw}

1371:   S.~Dimopoulos and G.~Landsberg,

1372: %  ``Black holes at the LHC,''

1373:   Phys.\ Rev.\ Lett.\  {\bf 87} (2001) 161602, hep--ph/0106295.

1374: 

1375: \bibitem{Eardley:2002re}

1376:   D.~M.~Eardley and S.~B.~Giddings,

1377: %  ``Classical black hole production in high-energy collisions,''

1378:   Phys.\ Rev.\ D {\bf 66} (2002) 044011, gr--qc/0201034.

1379: 

1380: \bibitem{Solodukhin:2002ui}

1381:   S.~N.~Solodukhin,

1382: %  ``Classical and quantum cross-section for black hole production in particle

1383: %  collisions,''

1384:   Phys.\ Lett.\ B {\bf 533} (2002) 153, hep--ph/0201248.

1385:   

1386: \bibitem{Hsu:2002bd}

1387:   S.~D.~H.~Hsu,

1388: %  ``Quantum production of black holes,''

1389:   Phys.\ Lett.\ B {\bf 555} (2003) 92, hep--ph/0203154.

1390:   

1391: \bibitem{Yoshino:2002tx}

1392:   H.~Yoshino and Y.~Nambu,

1393: %  ``Black hole formation in the grazing collision of high-energy particles,''

1394:   Phys.\ Rev.\ D {\bf 67} (2003) 024009, gr--qc/0209003.

1395: 

1396: \bibitem{Yoshino:2005hi}

1397:   H.~Yoshino and V.~S.~Rychkov,

1398: %  ``Improved analysis of black hole formation in high-energy particle

1399: %  collisions,''

1400:   Phys.\ Rev.\ D {\bf 71} (2005) 104028, hep--th/0503171.

1401: 

1402: \bibitem{Myers:1986un}

1403:   R.~C.~Myers and M.~J.~Perry,

1404: %  ``Black Holes In Higher Dimensional Space-Times,''

1405:   Annals Phys.\  {\bf 172} (1986) 304.

1406: 

1407: \bibitem{Hawking:1974sw}

1408:   S.~W.~Hawking,

1409: %  ``Particle Creation By Black Holes,''

1410:   Commun.\ Math.\ Phys.\  {\bf 43} (1975) 199,

1411:   [Erratum-ibid.\  {\bf 46} (1976) 206].

1412: 

1413: \bibitem{Emparan:2000rs}

1414:   R.~Emparan, G.~T.~Horowitz and R.~C.~Myers,

1415: %  ``Black holes radiate mainly on the brane,''

1416:   Phys.\ Rev.\ Lett.\  {\bf 85} (2000) 499, hep--th/0003118.

1417: 

1418: \bibitem{Anchordoqui:2002cp}

1419: L.~Anchordoqui and H.~Goldberg,

1420: %``Black hole chromosphere at the LHC,''

1421: Phys.\ Rev.\ D {\bf 67}, 064010 (2003), hep--ph/0209337.

1422: 

1423: \bibitem{charybdis}

1424: C.M. Harris, P. Richardson and B.R. Webber, JHEP {\bf 0308} (2003) 033,

1425: hep--ph/0307305.

1426: 

1427: \bibitem{catfish}

1428: M. Cavaglia, R. Godang, L. Cremaldi and D. Summers, hep--ph/0609001.

1429:  

1430: \bibitem{Dimopoulos:2001qe}

1431:   S.~Dimopoulos and R.~Emparan,

1432: %  ``String balls at the LHC and beyond,''

1433:   Phys.\ Lett.\ B {\bf 526} (2002) 393, hep-ph/0108060;

1434: K. Oda and N. Okada, Phys. Rev. {\bf D66} (2002) 095005, hep--ph/0111298.

1435: 

1436: \bibitem{raffelt}

1437: S. Hannestad and G.G. Raffelt, Phys. Rev. D {\bf 67} (2003) 125008

1438: [Erratum-ibid. D {\bf 69} (2004) 029901], hep--ph/0304029.

1439: %``Supernova and Neutron Star Limits on Large Extra Dimensions Reexamined.''

1440: 

1441: \bibitem{pdg}

1442: Particle Data Group, W.-M. Yao et al., Journal of Physics G {\bf 33} (2006) 1.

1443: 

1444: \bibitem{anchor3}

1445: L.A. Anchordoqui, J.L. Feng, H. Goldberg and A.D. Shapere, Phys. Rev. D {\bf

1446:   68} (2003) 104025, hep--ph/0307228.

1447: %``Updated Limits on TeV-Scale Gravity from Absence of Neutrino Cosmic Ray

1448: % Showers Mediated by Black Holes''

1449: 

1450: \bibitem{anchor2}

1451: L.A. Anchordoqui, J.L. Feng, H. Goldberg and A.D. Shapere, Phys. Lett. B {\bf

1452:   594} (2004) 363, hep--ph/0311365.

1453: %``Inelastic Black Hole Production and Large Extra Dimensions''

1454: 

1455: \bibitem{hoss1}

1456: S. Hossenfelder, Phys. Lett. B {\bf 598} (2004) 92, hep--th/0404232.

1457: %``Suppressed Black Hole Production from Minimal Length''

1458: 

1459: \bibitem{Ida1}

1460:  D.~Ida, K.~Oda and S.~C.~Park,

1461: %``Rotating black holes at future colliders: Greybody factors for brane

1462: % fields,''

1463: Phys.\ Rev.\ D {\bf 67} (2003) 064025, [Erratum-ibid.\ D {\bf 69} (2004)

1464:   049901], hep--th/0212108; and

1465: %  ``Rotating black holes at future colliders. II: Anisotropic scalar field

1466: %  emission,''

1467: Phys.\ Rev.\ D {\bf 71} (2005) 124039, hep--th/0503052.

1468: 

1469: \bibitem{kanti}

1470: P. Kanti and J. March-Russell, Phys. Rev. D {\bf 66} (2002) 024023,

1471: hep--ph/0203223, and Phys. Rev. D {\bf 67} (2003) 104019, hep--ph/0212199;

1472: C.M. Harris and P. Kanti, JHEP {\bf 0310} (2003) 014, hep--ph/0309054, and

1473: Phys. Lett. B {\bf 633} (2006) 106, hep--th/0503010; 

1474: G. Duffy, C. Harris, P. Kanti and E. Winstanley, JHEP {\bf 0509} (2005) 049,

1475: hep--th/0507274; 

1476: M. Casals, P. Kanti and E. Winstanley, JHEP {\bf 0602} (2006) 051,

1477: hep--th/0511163;

1478: M. Casals, S.R. Dolan, P. Kanti and E. Winstanley, hep--th/0608193.

1479: 

1480: \bibitem{Ida:2006tf}

1481: D.~Ida, K.~Oda and S.~C.~Park, Phys. Rev. D {\bf 73} (2006) 124022,

1482: hep--th/0602188. 

1483: 

1484: \bibitem{enhance}

1485: V. Frolov and D. Stojkovic, Phys. Rev. D. {\bf 66} (2002) 084002,

1486: hep--th/0206046, and Phys. Rev. Lett. {\bf 89} (2002) 151302, hep--th/0208102;

1487: M. Cavaglia, Phys. Lett. B {\bf 569} (2003) 7, hep--ph/0305256;

1488: D. Stojkovic, Phys. Rev. Lett. {\bf 94} (2005) 011603, hep--ph/0409124.

1489: 

1490: \bibitem{remnant}

1491: A. Bonanno and M. Reuter, Phys. Rev. D {\bf 62} (2000) 043008, hep--th/0002196,

1492: and Phys. Rev. D {\bf 73} (2006) 083005, hep--th/0602159;

1493: S. Hossenfelder et al., Phys. Lett. B {\bf 575} (2003) 85, hep--th/0305262.

1494: 

1495: \bibitem{ua2}

1496: UA2 collab., J. Alitti et al., Phys. Lett. B {\bf 268} (1991) 145.

1497: 

1498: \bibitem{cdf}

1499: CDF collab., F. Abe et al., Phys. Rev. D {\bf 56} (1997) 2532.

1500: 

1501: \bibitem{order}

1502: G. Marchesini and B.R. Webber, Nucl. Phys. B {\bf 238} (1984) 1.

1503: 

1504: \bibitem{Geiger:1998fq}

1505: K.~Geiger, R.~Longacre and D.~K.~Srivastava, nucl--th/9806102.

1506: %VNI version 4.1: Simulation of high-energy particle collisions in QCD

1507: %Space-time evolution of e+ e- $\to$ A + B collisions with parton  cascades,

1508: %parton hadron conversion, final-state hadron cascades,''

1509: 

1510: \bibitem{Bass:1999pv}

1511: S.~A.~Bass, M.~Hofmann, M.~Bleicher, L.~Bravina, E.~Zabrodin, H.~Stoecker and

1512: W.~Greiner, 

1513: %``Analysis of reaction dynamics at ultrarelativistic energies in a combined

1514: %parton hadron transport approach'',

1515: Phys.\ Rev.\ C {\bf 60} (1999) 021901, nucl--th/9902055. The code is available

1516: from http://www.phy.duke.edu/~bass/vni.html .

1517: 

1518: \bibitem{Sjostrand:2003wg}

1519: T.~Sj\"ostrand, L.~Lonnblad, S.~Mrenna and P.~Skands, hep--ph/0308153.

1520: %``PYTHIA 6.3: Physics and manual,''

1521: 

1522: \bibitem{gnu}

1523: See http://www.gnu.org/software/gsl.

1524: 

1525: \bibitem{Geiger:1991nj}

1526: K.~Geiger and B.~M\"uller,

1527: %``Dynamics of parton cascades in highly relativistic nuclear collisions,''

1528: Nucl.\ Phys.\ B {\bf 369} (1992) 600.

1529: 

1530: \bibitem{Bengtsson:1986et}

1531: M.~Bengtsson and T.~Sj\"ostrand,

1532: %``A Comparative Study Of Coherent And Noncoherent Parton Shower Evolution,''

1533: Nucl.\ Phys.\ B {\bf 289} (187) 810.

1534: 

1535: \bibitem{Webber:1986mc}

1536: B.~R.~Webber,

1537: %``Monte Carlo Simulation Of Hard Hadronic Processes,''

1538: Ann.\ Rev.\ Nucl.\ Part.\ Sci.\  {\bf 36} (1986) 253.

1539: 

1540: \bibitem{schwinger}

1541: See e.g. J. Schwinger, ``Particles, Sources and Fields'', Vol. II

1542: (Addison--Wesley, Reading, MA, 1970/73).

1543: 

1544: \end{thebibliography}

1545: \end{document}

1546: