1: \documentstyle[twocolumn,prc,aps,epsf,epsfig]{revtex}

2: 

3: \newcommand{\po}{\partial_\omega}

4: \newcommand{\pt}{\partial_\tau}

5: \newcommand{\dn}{{\rm dn}}

6: \newcommand{\ve}{\varepsilon}

7: \newcommand{\IA}{$\bar I I \,$}

8: \begin{document}

9: \twocolumn[\hsize\textwidth\columnwidth\hsize\csname@twocolumnfalse\endcsname

10: \preprint{SUNY-NTG-xxx}

11: \preprint{NT@UW-02-008}

12: 

13: \title{

14: Forced Tunneling and Turning State Explosion in Pure Yang-Mills Theory 

15: }

16: 

17: \author{D. M. Ostrovsky$^1$, G. W. Carter$^2$, and E. V. Shuryak$^1$ }

18: 

19: \address{

20: $^1$Department of Physics and Astronomy, State University of 

21: New York, Stony Brook, NY 11794-3800\\

22: $^2$Department of Physics, Box 351560, University of Washington, Seattle,

23: WA 98195-1560

24: }

25: 

26: \date{18 April 2002}

27: \maketitle

28: 

29: \begin{abstract}

30: We consider {\em forced} tunneling in QCD, described semiclassically by

31: instanton-antiinstanton field configurations.

32: By separating topologically different minima we obtain details of the 

33: effective potential and study the {\em turning states}, which are similar

34: to the sphaleron solution in electroweak theory.

35: These states are alternatively derived as minima of the

36: energy under the constraints of fixed size and Chern-Simons number.

37: We study, both analytically and numerically, the subsequent evolution of 

38: such states by solving the classical Yang-Mills equations in real time, 

39: and find that the gauge field strength is quickly localized into an

40: expanding shell of radiating gluons.

41: The relevance to high-energy collisions of hadrons and

42: nuclei is briefly discussed.

43: \end{abstract}

44: \vskip 2pc]

45: 

46: \section{Introduction}

47: 

48: \subsection{Instanton-Induced Scattering in QCD}

49: 

50: The existence of topologically distinct non-abelian

51: gauge fields, with tunneling

52: between corresponding classical vacua described semiclassically by

53: instantons \cite{BPST}, is one of the most spectacular nonperturbative

54: effects of field theory.

55: Significant progress has been made in understanding instanton-induced

56: effects in Quantum Chromodynamics (QCD), 

57: explaining both explicit U$_A$(1) chiral symmetry

58: breaking at the single-instanton level \cite{tHooft} and spontaneous

59: SU($N_f$) chiral symmetry breaking by the instanton 

60: ensemble \cite{inst_chiral}. 

61: Euclidean correlation functions, studied phenomenologically and on the 

62: lattice, have been explained to a significant extent by instantons as well

63: \cite{SS_98}.

64: 

65: With tunneling phenomena apparently so important in {\em virtual}

66: quark and gluon propagation, it is reasonable to think them

67: also relevant in {\em real} processes such as scattering or 

68: particle production in Minkowski space.

69: We thus seek contributions to parton scattering amplitudes from 

70: the theory of instanton-related objects, and supporting experimental evidence.

71: 

72: With this as our motivation, we concentrate in this paper on the

73: theoretical basis of such effects from pure Yang-Mills theory.

74: Specific applications to high-energy processes with hadrons or nuclei

75: are left for papers to follow, although we will discuss phenomenological

76: generalities where relevant.

77: 

78: Progress in understanding of the role of tunneling in

79: high energy processes has been tempered by technical problems for years.

80: Significant insights were obtained in the 1980's \cite{electroweak80s}

81: and further developed in the early 1990's \cite{electroweak90s,KhR} 

82: through work in electroweak theory.

83: In this case, the instanton-induced cross section is readily identified by 

84: baryon number violation and many noteworthy features of these processes 

85: were found.

86: However, quantitative estimates of the associated cross sections

87: proved to be far below observable limits and interest quickly waned.

88: Similar ideas have also been developed in QCD \cite{MRS}, 

89: notably the search for hard 

90: processes induced by {\em small-sized instantons} which continues

91: at HERA \cite{RS}.

92: 

93: Another role for instanton-induced processes has recently been proposed by

94: Kharzeev, Kovchegov, and Levin \cite{KKL} and

95: Nowak, Shuryak, and Zahed \cite{NSZ}.

96: These works focus on typical QCD instantons, of size $\rho\sim 1/3$ fm

97: \cite{inst_chiral}, which determine the {\em semi-hard} scale 

98: of $Q\sim 1-2$ GeV. 

99: It was proposed  that topological tunneling is behind the well-known 

100: features of high energy scattering described phenomenologically by the 

101: so-called ``soft'' pomeron.

102: These ideas were further tested in Ref.~\cite{COS}, where they were

103: demonstrated to be reasonably consistent with experimental data.

104: 

105: Since the 1960's attempts have been made to explain high-energy hadronic

106: collisions with multi-peripheral models, with various ladder diagrams

107: describing hadron production. 

108: It was realized that in order to get cross-sections which are not falling at 

109: high energies, one needed $vector$ field exchange in the $t$-channel. 

110: With the discovery of QCD, gluons naturally play this role. 

111: Generic pQCD-inspired models appeared with processes like that

112: shown in Fig.~\ref{processes}(a).

113: Eventually this development led to the BFKL gluon ladder \cite{BFKL}, 

114: which produces an (approximately) supercritical pomeron, a

115: ``hard'' pomeron with the intercept well above 1. 

116: Recent studies of high energy hard processes, especially at HERA, 

117: have indeed found strong growth of the cross section with energy for truly 

118: hard processes ($Q^2 \gg 1$ GeV$^2$), consistent with the BFKL

119: treatment.

120: 

121: But various data at the {\em semi-hard} scale of $Q^2 \sim 1$ GeV$^2$

122: demonstrate rather different growth with energy, consistent with a

123: ``soft'' pomeron.

124: Whatever it might be, the pomeron should be an object of a particular size

125: deduced from the slope of its Regge trajectory,

126: $\alpha'\sim 1/(2 \, {\rm GeV})^2$. 

127: This size of course cannot be explained by basically scale-invariant pQCD,

128: and thus calls for a nonperturbative derivation.

129: 

130: Existing models for the soft pomeron also include ladders made of $t$-channel

131: gluons, and the differences between them lie mainly in the construction of

132: their rungs. 

133: Each of the various models has a unique answer for

134: {\em what is actually produced} in gluon-gluon partonic collisions.

135: For example, in Ref.~\cite{KL} a pair of pions in the scalar channel 

136: or a scalar glueball is produced.

137: The introduction into this problem of instanton-induced vertices

138: \cite{Shu_toward,KKL,NSZ}, shown schematically in Fig.~\ref{processes}(b),

139: led to a different idea: the object produced is neither a gluon (as in BFKL)  

140: nor any {\em colorless} hadronic state, but rather a {\em colored} cluster 

141: of the gluon field, which in turn decays into several gluons. 

142: It has been shown that the cross section peaks at an invariant cluster mass

143: in the range $2.5-3$ GeV \cite{KKL,NSZ}.

144: It is very important that the states which are produced are not a random

145: group of gluons, but rather their coherent superposition.

146: Understanding their composition is the main objective of this work.

147: 

148: A quantum-mechanical interpretation of the collision process is central 

149: to this question of prompt gluon production.  

150: An impressive body of work has addressed this problem with classical 

151: Weizs\"acker-Williams fields of gluons, the Color Glass Condensate

152: \cite{cgc}.

153: Here we consider a different classical process, one involving topological

154: objects.

155: In Fig.~\ref{processes}(c) we schematically show a barrier separating 

156: two topologically distinct classical vacua, with

157: Chern-Simons numbers\footnote{This will be introduced formally below. Here it

158: is sufficient to note only that we consider a definite pair of gauge

159: potentials, separated on one of the many coordinates of our quantum system.} 

160: $N_{CS}=$ 0 and 1. 

161: Unlike a standard instanton transition, shown by the horizontal dashed line, 

162: in a high energy collision a finite amount of energy is absorbed.

163: This can be viewed as a ``forced tunneling'' event (either of the other 

164: two dashed lines) which ends at a {\em turning state}, where the total 

165: energy is equal to the potential energy, so that the paths can exit the 

166: (Euclidean) domain below the barrier.

167: These colored unstable objects are close relatives of electroweak 

168: sphalerons \cite{Manton,KM}, which are defined at the barrier's peak. 

169: We will demonstrate how these objects then evolve with conserved energy, 

170: developing into an exploding shell of color field.

171: This part of the process is diagrammed with the horizontal lines in

172: Fig.~\ref{processes}(c).

173: 

174: Before we come to these explosions, we will discuss in detail the

175: instanton--anti-instanton (\IA) configurations which describe this

176: forced tunneling. 

177: They provide one way toward the understanding of the 

178: effective potential separating topologically different gauge fields, 

179: as well as the turning states themselves.

180: We then proceed to another derivation of the same results as static 

181: solutions in classical Yang-Mills theory constrained in size.

182: The real-time decay of the static configuration is studied in detail, using 

183: both analytic and numerical methods, ultimately leading to a description

184: of the expanding shells in terms of gluonic quanta.

185: %########################################################################

186: \begin{figure}[ht]

187: \begin{center}

188:   \begin{minipage}[c]{1.8in}

189:     \centering

190:      \includegraphics[totalheight=1.8in]{fig1.eps}

191:    \end{minipage}

192: \vskip 5mm 

193:   \begin{minipage}[c]{2.6in}

194:     \centering

195:       \includegraphics[width=2.6in]{fig2.eps}

196:    \end{minipage}

197: \end{center}

198:   \caption[]

199:   {

200:    \label{processes}

201: The top of the figure compares

202: (a) a typical inelastic perturbative process (two t-channel gluons collide,

203: producing

204: a pair of gluons) to (b) a nonperturbative inelastic process, incorporating

205: collisions

206: of a few $t$-channel gluons with the instanton (the shaded circle),

207: resulting in multi-gluon production.

208: The bottom figure (c) shows the same process, but in a 

209: quantum mechanical way.  The energy of Yang-Mills field versus

210: the Chern-Simons number, $N_{cs}$, is a periodic

211: function, with zeros at integer points. The {\em instanton} (shown by the 

212: lowest dashed line) is a transition between such points. However if some nonzero

213: energy is deposited into the process during transition, the virtual path

214: (the dashed line) leads to a  {\em turning states}, from which starts the

215: real time motion outside the barrier (shown by horizontal solid lines).

216: The maximal cross section corresponds to the transition to

217: the top of the barrier, called a {\em sphaleron}.

218:  }

219: \end{figure}

220: %########################################################################

221: 

222: \subsection{Spherically Symmetric Yang-Mills Fields}

223: 

224: For the SU(2) color subgroup in which we are interested, spherically 

225: symmetric field

226: configurations of the gauge field ${\cal A}_\mu^a$ can be expressed 

227: through the following four space-time (0, $j=1..3$) 

228: and color ($a=1..3$) structures 

229: \begin{eqnarray}

230: \label{sph_ansatz}

231: {\cal A}^a_j &=& A(r,t)\Theta^a_j + B(r,t) \Pi^a_j + C(r,t)\Sigma^a_j

232: \nonumber\\

233: {\cal A}^a_0 &=& D(r,t) \frac{x^a}{r}

234: \end{eqnarray}

235: with

236: \begin{equation}

237: \Theta^a_j=\frac{\epsilon_{jam}x^m}{r}\,, \quad \Pi^a_j =

238: \delta_{aj}-\frac{x_ax_j}{r^2}\,, \quad \Sigma^a_j = \frac{x_ax_j}{r^2}\,.

239: \label{projectionops}

240: \end{equation}

241: It is convenient to express the scalar functions in Eq.~(\ref{sph_ansatz}) 

242: in terms of four $r$ and $t$ dependent functions, which are similar to the 

243: fields of the 1+1 dimensional Abelian gauge-Higgs model

244: ($A_{\mu=0,1},\,\phi,\, \alpha$)

245: on a hyperboloid \cite{Witten_ans}:

246: \begin{equation}

247: A=\frac{1+\phi\sin\alpha}{r}, \quad B=\frac{\phi\cos\alpha}{r},\quad

248: C=A_1,\quad D=A_0.

249: \end{equation}

250: 

251: One can express the field strengths in these terms as

252: \begin{eqnarray}

253: {\cal E}^a_j = {\cal G}^a_{0j} 

254: &=&\frac{1}{r} [\partial_0\phi\sin\alpha +

255: \phi\cos\alpha(\partial_0\alpha-A_0)]\Theta^a_j \nonumber\\

256: &&+ \frac{1}{r} [\partial_0\phi\cos\alpha -

257: \phi\sin\alpha(\partial_0\alpha-A_0)]\Pi^a_j

258: \nonumber\\&&

259: + (\partial_0A_1-\partial_1A_0)\Sigma^a_j

260: \label{electric}\end{eqnarray}

261: and

262: \begin{eqnarray}

263: {\cal B}^a_j = \frac{1}{2}\epsilon_{jkl}{\cal G}^a_{kl} 

264: &=&\frac{1}{r} [-\partial_1\phi\cos\alpha +

265: \phi\sin\alpha(\partial_1\alpha-A_1)]\Theta^a_j \nonumber\\

266: &&+ \frac{1}{r} [\partial_1\phi\sin\alpha +

267: \phi\cos\alpha(\partial_1\alpha-A_1)]\Pi^a_j 

268: \nonumber\\ &&

269: + \frac{1-\phi^2}{r^2}\Sigma^a_j, 

270: \label{magnetic}\end{eqnarray}

271: where $\partial_0\equiv \partial_t \mbox{ and } \partial_1\equiv\partial_r$.

272: 

273: The action in 3+1 dimensional Minkowski $(-,+,+,+)$ space reduces as

274: \begin{eqnarray}\label{action}

275: S &=& \frac{1}{4g^2}\int d^3x dt\left[\left({\cal B}^a_j\right)^2 

276: - \left({\cal E}^a_j\right)^2\right]

277: \nonumber\\

278: &=& 4\pi\int dr dt \Bigg[\left(\partial_\mu\phi\right)^2+

279: \phi^2\left(\partial_\mu\alpha-A_\mu\right)^2\nonumber\\

280: &&+\frac{(1-\phi^2)^2}{2r^2}-

281: \frac{r^2}{2}\left(\partial_0A_1-\partial_1A_0\right)^2\Bigg] \,,

282: \end{eqnarray}

283: with the summation now over the 1+1 dimensional $(-,+)$ metric.

284: 

285: The spherical ansatz is preserved by a set of gauge transformations

286: generated by unitary matrices of the type

287: \begin{equation}

288: \label{gauge_transform}

289: U(r,t)=\exp\left(i\frac{\beta(r,t)}{2r}\tau^ax^a\right).

290: \end{equation}

291: These transformations naturally coincide with the gauge symmetry of 

292: the corresponding abelian Higgs model:

293: \begin{equation}

294: \phi'=\phi, \quad \alpha'=\alpha+\beta, \quad A'_\mu=A_\mu+\partial_\mu

295: \beta\,.

296: \end{equation}

297: This freedom can be used to gauge out, for example, the 

298: ${\cal A}_0$ component.

299: 

300: Topological properties of the gauge field are governed by the

301: topological current

302: \begin{equation}

303: K_\mu=-{1 \over 32 \pi^2} \epsilon^{\mu\nu\rho\sigma}

304: \left({\cal G}^a_{\nu\rho} {\cal A}^a_\sigma -

305: {g \over 3}\epsilon^{abc} {\cal A}^a_\nu  {\cal A}^b_\rho  {\cal A}^c_\sigma

306: \right)\,.

307: \end{equation}

308: Although this current is not gauge invariant, its change is

309: related to the (gauge invariant) local topological charge

310: \begin{equation}

311: \partial_\mu K^\mu

312: = -{1 \over 32 \pi^2} {\cal G}^a_{\mu\nu}\tilde {\cal G}^a_{\mu\nu} \,.

313: \end{equation}

314: Within the spherical ansatz and the ${\cal A}_0=0$ gauge the topological

315: current takes a simpler form,

316: \begin{eqnarray}

317: K^0 &=& \frac{1}{8\pi^2r^2}

318: \left[(1-\phi^2)(\partial_1\alpha-A_1)-\partial_1(\alpha-\phi\cos\alpha)\right]

319: \nonumber\\

320: K^i &=& \frac{x^i}{8\pi^2r^3}

321: \left[(1-\phi^2)\partial_0\alpha-\partial_0(\alpha-\phi\cos\alpha)\right] \,,

322: \end{eqnarray}

323: while the topological charge becomes

324: \begin{eqnarray}

325: \partial_\mu

326: K^\mu &=& \frac{1}{8\pi^2r^2}\left\{-\partial_0\left[(1-\phi^2)

327: (\partial_1\alpha-A_1)\right]\right.

328: \nonumber\\

329: &&+\left.\partial_1\left[(1-\phi^2)(\partial_0\alpha-A_0)\right]\right\}\,.

330: \end{eqnarray}

331: Note that only gauge-invariant combinations of field derivatives 

332: appear here.

333: 

334: As a ``topological coordinate'' marking the tunneling paths and

335: the turning states one can use the Chern-Simons number

336: \begin{eqnarray}

337: N_{CS} = \int d^3x K_0

338: &=& -\frac{1}{2\pi}\int dr

339: (1-\phi^2)(\partial_1\alpha-A_1)

340: \nonumber\\&&

341: +\frac{1}{2\pi} \left.(\alpha-\cos\alpha)\right|_{r=0}^{r=\infty}

342: \label{csdef}

343: \end{eqnarray}

344: The first, gauge-invariant term is sometimes called

345: the {\em corrected} or {\em true} Chern-Simons number \cite{AKY,FKS},

346: $\tilde{N}_{CS}$, while the second (gauge-dependent) term is referred to as

347: the {\em winding number}. It is the change in $\tilde{N}_{CS}$ which

348: is equivalent to the integral over the local topological charge.

349: 

350: 

351: \section{ Instanton-Antiinstanton Configurations}

352: 

353: \subsection{Forced Tunneling}

354: 

355: A brief introduction to the quantum mechanics of gluons in high energy

356: collisions has been given in the introduction.

357: The effect of colliding partons can be included in various forms.

358: For example, these fields can be represented as {\em non-zero}

359: external currents which affect the tunneling paths of Yang-Mills field. 

360: In the zero-current, vacuum case, the usual instanton solutions

361: are spherically symmetric in four Euclidean dimensions. 

362: The collision problem of two (or more) partons, on the contrary, at

363: non-zero impact parameters does not have even an {\em axial} symmetry.

364: The reader therefore may wonder why this (and all previous works)

365: on the subject consider 3+1 dimensional spherically symmetric fields.

366: 

367: The justification for this ansatz is that the absolute magnitude of the 

368: tunneling field is large compared to external forces.

369: Also, as will be shown below, 

370: spherically symmetric clusters are an energy minimum

371: for fixed size and topological coordinate. 

372: Should the resulting cluster not have exact spherical symmetry

373: one can always approach the problem perturbatively, considering first the 

374: external forces projected onto the direction of tunneling, and then other

375: components as small corrections.

376: The resulting 1+1 dimensional problem is readily solved numerically and,

377: to a great extent, analytically.

378: 

379: Unlike separated instantons ($I$) and antiinstantons ($\bar I$), 

380: combined \IA configurations are neither selfdual nor anti-selfdual

381: and do not satisfy classical equations of motion.  

382: They are not extrema of the action, since they describe the valley 

383: stretching between true extrema -- the zero field (equivalent to

384: an \IA at zero separation) and well-separated \IA pair.

385: Substituting any \IA trial function into the of Yang-Mills equation

386: of motion, we find a finite

387: \begin{equation} 

388: D_\mu {\cal G}_{\mu\nu}=J_\nu  \,.

389: \end{equation} 

390: This means some external current must be applied to the gauge fields if we 

391: want to use semiclassical analysis.

392: The process can only then be interpreted as a classical \IA, or a

393: {\em forced path}.

394: There are two interpretations of \IA configurations with different

395: consequences.

396: 

397: The historical view is that such fields describe quantum fluctuations in 

398: the Yang-Mills vacuum, the process in which a virtual path goes

399: under the barrier, then reverses course and ends up in the {\em same} 

400: minimum from which it started.

401: This process has zero net topological charge.

402: Naturally, the early studies concentrated on the action corresponding 

403: to these configurations, the quantity which controls its weight in the 

404: path integral.

405: The first such work was done long ago by Callan, Dashen, and Gross

406: \cite{CDG}, resulting in a dipole force and the

407: action $\delta S \sim 1/T^4$ at large distance $T$ between the centers.

408: Higher terms in the multipole expansion have been discussed

409: in literature after that, e.g. \cite{yung}. 

410: When it was eventually realized that quark-induced \IA pairings are more 

411: important for the instanton ensemble in QCD \cite{SS96}, interest in the 

412: pure Yang-Mills theory waned.

413: 

414: In this paper we will however take a different view of \IA configurations.

415: Since the external forces from the partonic current do work on the \IA

416: pair, the energy at intermediate times is non-zero. 

417: We will consider only cases in which the fields at positive and negative times

418: are essentially the same (modulo a sign and, sometimes, a

419: gauge transformation).  

420: Thus this energy will be even under $t\rightarrow -t$, with a

421: natural maximum at $t=0$. 

422: As a result, all quantities which are odd under this transformation

423: (like the electric field) naturally vanish at this instant.

424: The remaining, {\em purely magnetic} configuration

425: is what we define as the {\em turning state} of this path.

426: 

427: The resulting action corresponds to an excitation {\em probability}

428: of this turning state created by the external current $J$,

429: \begin{equation} 

430: P \sim \left| \langle 0|J|{\rm turning \, state}\rangle \right|^2 .

431: \end{equation} 

432: Through this mechanism the excitation of \IA pairs leads to the production

433: of real particles, as advertized in the Introduction and to be analyzed 

434: in the next sections.

435: 

436: \subsection{Simple \IA Trial Functions}

437: 

438: We now consider the simplest example of a possible turning state, a

439: straightforward \IA sum ansatz.

440: With it, we can demonstrate some basic features, although we will find

441: them insufficient for our purposes and move to a more complicated

442: ansatz in the next subsection.

443: 

444: Written the singular gauge, the {\em sum ansatz} is:

445: \begin{equation} 

446: \frac{g}{2} {\cal A}_{a\mu}^{sum}(x)= 

447: \frac{ \bar \eta_{a\mu\nu} y_1^\nu \rho^2}{ y_1^2(y_1^2+\rho^2)}+ 

448: \frac{ \eta_{a\mu\nu} y_2^\nu \rho^2 }{y_2^2(y_2^2+\rho^2)}\,,

449: \end{equation} 

450: where we assume that both the instanton and the antiinstanton (the first 

451: and second terms, respectively) have the same color orientation and size

452: $\rho$. 

453: The vectors $y_1=x-z_I$ and $y_2=x-z_{\bar I}$ are the distances from

454: the observation point $x$ to the instanton and antiinstanton centers. 

455: In what follows we assume $z_I=(T/2,0,0,0)$ and $z_{\bar I}=(-T/2,0,0,0)$, 

456: where the imaginary time between centers is $T$.

457: 

458: Note that although a single instanton's profile behaves as $1/x$ near the

459: origin, the physical quantity $\left({\cal G}^a_{\mu\nu}\right)^2$ is finite.

460: However, for the sum ansatz this feature is lost and the same quantity 

461: goes as $1/x^2$ near the origin.

462: 

463: This unphysical feature can be quickly remedied by the {\em ratio 

464: ansatz} \cite{Shu_rat}, which for identical sizes and orientations is

465: \begin{equation} 

466: \frac{g}{2} {\cal A}_{a\mu}^{ratio}(x) = \frac{\eta_{a,\mu\nu}y_1^{\nu}

467: \frac{\rho^2}{y_1^2} +\bar\eta_{a,\mu\nu}y_2^{\nu}

468: \frac{\rho^2}{y_2^2} }{ 1+ \frac{\rho^2}{y_1^2}+\frac{\rho^2}{y_2^2} }

469: \end{equation} 

470: These trial functions are simple enough to have analytic

471: expressions for the field strength, the energy of static turning states,

472: and the Chern-Simons number. 

473: For reference, one has the following

474: expressions for the magnetic and electric fields squared:

475: \begin{eqnarray} 

476: \vec {\cal B}^2 &=& 16384(768 t^8 + 1024 r^2 t^6 + 3072 t^6 + 2304 t^6 R^2 

477: \nonumber\\ && 

478: + 6400 r^2 t^4 R^2 + 2048 r^2 t^4 + 512 r^4 t^4 + 1824 t^4 R^4

479: \nonumber \\ && 

480: + 3072 t^4 + 4608 t^4 R^2 + 1024 r^6 t^2 + 192 t^2 R^4

481: \nonumber \\ && 

482: + 512 r^2 t^2 R^2 - 1024 r^4 t^2 + 144 t^2 R^6 + 1216 r^2 t^2 R^4 

483: \nonumber \\ && 

484: + 2816 r^4 t^2 R^2 + 288 r^4 R^4 + 768 r^8 + 768 r^6 R^2 

485: \nonumber \\ && 

486: + 48 r^2 R^6 + 3 R^8)/(16 r^4 + 32 r^2 t^2 + 8 r^2 R^2 + 16 t^4 

487: \nonumber \\ && 

488: - 8 t^2 R^2 + R^4 + 32 r^2 + 32 t^2 + 8 R^2)^4 \,,

489: \end{eqnarray}

490: \begin{eqnarray} 

491: \vec {\cal E}^2 &=& 1048576 t^{2}(32 r^2 t^4 + 48 t^4 R^2

492: + 64 r^4 t^2 + 64 r^2 t^2 \nonumber \\ 

493: && + 80 r^2 t^2 R^2

494: + 48 t^2 R^2 + 24 t^2 R^4 + 12 R^4 + 32 r^2 \nonumber \\  

495: && + 32 r^6 + 64 r^2 R^2 + 26 r^2 R^4 + 64 r^4 R^2 \nonumber \\

496: && + 64 r^4 + 12 R^2 + 3 R^6)/

497: %% \nonumber \\ && 

498: (16 r^4 + 32 r^2 t^2 + 8 r^2 R^2 \nonumber \\ 

499: && + 16 t^4 - 8 t^2 R^2 + R^4 + 32 r^2 + 32 t^2 + 8 R^2)^4 \,.

500: \end{eqnarray} 

501: Their scalar product is

502: \begin{eqnarray} 

503: \vec {\cal B}\cdot\vec {\cal E} &=& -393216 t R(R^2 + 2 + 4 r^2 + 4 t^2)(16 t^4 

504: \nonumber \\ && 

505: + 24 t^2 R^2 + 32 r^2 t^2 + 32 t^2 + R^4 + 16 r^4 

506: \nonumber \\ && 

507: + 8 r^2 R^2)/ (16 r^4 + 32 r^2 t^2 + 8 r^2 R^2 + 16 t^4 

508: \nonumber \\ && 

509: - 8 t^2 R^2 + R^4 + 32 r^2 + 32 t^2 + 8 R^2)^4 \,,

510: \end{eqnarray} 

511: where we have set $\rho=1$ and $R=T$ is the intercenter distance. 

512: 

513: One can see that, in the simplest case of identical sizes and orientations

514: for the $I $ and $\bar I$, time reflection symmetry $t\rightarrow -t$ 

515: of the problem is indeed manifest, so that

516: \begin{equation} 

517: {\cal A}^a_0(\vec r,t=0)=0 \,,\quad

518: {\cal E}^a_m(\vec x,t=0)=0\,. 

519: \end{equation} 

520: This is illustrated in Fig.~\ref{IAfig}(b).

521: Since configurations of this type interpolate between a mostly dual region, with

522: ${\cal E}^a_m(z_I)={\cal B}^a_m(z_I)$, to an anti-dual region, 

523: where ${\cal E}^a_m(z_{\bar I})=-{\cal B}^a_m(z_{\bar I})$, it is intuitive

524: that the electric field vanishes in the center.

525: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

526: \begin{figure}[tb]

527: \hspace*{-5mm}

528: \begin{center}

529: \epsfig{file=fig3.eps, width=45mm}

530: \vskip 5mm 

531: \epsfig{file=fig4.eps, width=50mm}

532: \end{center}

533: \caption {\label{IAfig}

534: Instanton-antiinstanton configurations. (a) A schematic picture in 

535: Euclidean space-time. The thick vertical line, $t=0$, corresponds to the

536: location of the turning state. The definition of

537: the inter-center distance $T$ is also shown. 

538: (b) Distribution along the time axis

539: of $2\vec {\cal B}^2$,$2\vec {\cal E}^2$, and 

540: $2\vec {\cal B}\cdot \vec {\cal E}$ 

541: for the ratio ansatz with $T=\rho$,

542: shown by the solid, dashed, and short-dashed lines respectively.

543: The curve for $\vec {\cal B}\cdot\vec{\cal E}$ is the only one which is $t$-odd.

544: }

545: \end{figure}

546: 

547: This situation can be readily interpreted in the ${\cal A}_0=0$ gauge, 

548: in which the electric field is simply the time derivative of the

549: gauge field -- the canonical {\em momentum} in Yang-Mills field quantization.

550: Thus the $t=0$ magnetic state is indeed identified as 

551: a turning state, in which motion is momentarily stopped.

552: For separation $T$ comparable to the size $\rho$ the energy is finite,

553: with a maximum $E\sim 1/(g\rho)$.

554: 

555: The energy $E$ and Chern-Simons number $N_{CS}$ for either the sum or ratio 

556: ansatz can be calculated as a function of separation $T$ directly, 

557: with the hope that a parametric plot of $E(N_{CS})$ will reveal a useful

558: profile of the barrier as a function of this topological coordinate.

559: 

560: Alas, for the sum ansatz this idea produces reasonable results only for very 

561: large separation, $T\geq 2\rho$.

562: When $T$ is of the order $\rho$, the energy $E(T)$ of the turning state 

563: (as well as the action for the entire configuration) becomes very large, 

564: while the topological coordinate $N_{CS}(T)$ remains fixed. 

565: It is therefore obvious that this set of paths does not describe the travel 

566: across the ridge separating classical vacua which we want to study.

567: Instead, this path rises with the barrier but continues to increase as the

568: origin is approached, following a direction apparently orthogonal to

569: the topological coordinate we want to study.

570: 

571: The ratio ansatz yields somewhat better results, with finite (and

572: even simple) field structure at all $T$, including the point $T=0$. 

573: However the results, shown in Fig.~\ref{rat_path},

574: indicate that this set of trial functions can only accomplish about 

575: one third of the journey we would like to make, in terms of the topological 

576: quantity $N_{CS}$.

577: This inadequacy will become apparent after comparison with the results 

578: to follow.

579: \begin{figure}[h]

580: \hspace*{-5mm}

581: \begin{center}

582: \epsfig{file=fig5.eps, width=50mm}

583: \end{center}

584: \caption {\label{rat_path}

585: The normalized energy, $ER$, versus the Chern-Simons number

586: for the ratio ansatz.

587: }

588: \end{figure}

589: 

590: \subsection{The Yung Ansatz, or Going Uphill} 

591: 

592: As a natural set of \IA configurations, one of us \cite{Shu_qm}

593: suggested starting from a well-separated pair and 

594: going {\em downhill}, along the gradient of the action\footnote{

595: This can easily be done numerically, and a set of such curves

596: for the quantum-mechanical double well potential and

597: the corresponding set of \IA configurations was found in that work.}. 

598: Naturally, minimization of the action leads to

599: complete \IA annihilation and zero action.

600: 

601: It was shown by Yung that these configuration can generally be obtained 

602: from a solution of the streamline equation \cite{yung}. 

603: He found solutions for large separation $T\gg \rho$ and used them to derive

604: the next order terms in the \IA interaction, to $O(1/T^6)$. 

605: A clever conformal symmetry was used to reduce the

606: Yang-Mills problem to that of a double well potential.

607: The same trick was then used in the numerical solution of the streamline 

608: equation \cite{KhR,Verbaarschot}, in which it was observed

609: that the approximate ansatz suggested by Yung also happens to be  

610: a very accurate approximation to true solution, not only at large $T$

611: (as Yung intended) but in fact for {\em all} finite \IA separations $T$. 

612: As expected, at $T=0$ \IA annihilation occurs and

613: the field strength vanishes\footnote{

614: This is not obvious from the Yung expression; it was first found numerically.

615: The Yung formula's complicated result at $T=0$ is nothing but a 

616: pure gauge.}.

617: 

618: Since we take a different view of \IA configurations in this work,

619: we interpret a solution of the streamline equation (or Yung ansatz) as

620: a set of forced paths going {\em uphill} against the gradient of the

621: force. 

622: This process reaches its turning point (or state), with some maximal energy

623: and Chern-Simons number, and then turns back.  

624: Because the process proceeds uphill, unlike with other trial functions

625: with some arbitrary driving force, we expect that all trajectories rise 

626: along {\em the same path}, although those with larger $T$ go further up.

627: 

628: The Yung ansatz for the field configuration is rather complicated,

629: and is best written in matrix form: 

630: \begin{eqnarray} 

631: ig{\cal A}_{\mu}^{Yung}(x) &=& 

632: ig{\cal A}_{a\mu}^{Yung}(x)\frac{\tau^a}{2} \nonumber \\

633: &=& {\bar {\tilde y_2} \over \sqrt{\tilde y_2} }

634:  {R \over \sqrt{R^2} } 

635:  {(\bar \sigma_\mu y_1-y_1^\mu)\rho_1^2 \over y_1^2 ( y_1^2+\rho_1^2)}

636:  {\bar R \over\sqrt{R^2}}{\bar {\tilde y_2} \over \sqrt{\tilde y_2}}  

637:  \nonumber \\

638: && + {(\bar \sigma_\mu y_2-y_2^\mu)\rho_2^2 \over  y_2^2+\rho_2^2}+ 

639:  {\rho_1\rho_2 \over z y_1^2( y_2^2+\rho_2^2)}\nonumber \\

640: && \Bigg[ (\bar \sigma_\mu y_1-y_1^\mu)-

641: {\bar {\tilde y_2} \over \sqrt{\tilde y_2}}\nonumber \\

642: && {R \over \sqrt{R^2}} (\bar \sigma_\mu y_1-y_1^\mu) 

643: {\bar R \over\sqrt{R^2}}{\bar {\tilde y_2} \over \sqrt{\tilde y_2}}

644: \Bigg] ,

645: \end{eqnarray} 

646: where $z$ is related to the conformal-invariant distance,

647: $(R^2+\rho_1^2+\rho_2^2)/(\rho_1\rho_2)$. 

648: In the case

649: $\rho_1=\rho_2=\rho$, which is the only one we need, this relation reads

650: \begin{equation} 

651: z^2=\frac{R^2+2\rho^2+\sqrt{(R^2+2\rho^2)^2-4\rho^2} }{2 \rho^2} \,.

652: \end{equation} 

653: All vectors without an indicative index are SU(2) matrices

654: obtained by their contraction with the vector $\sigma_\mu=(1,-i\vec\tau)$, 

655: for example $R=x_1-x_2=R_\mu \sigma_\mu$.  

656: An overbar similarly denotes contraction with $\bar\sigma= (1,i\vec\tau)$. 

657: Note that barred and unbarred matrices always alternate, in all terms; 

658: this is because one index of each matrix is dotted and the other not, 

659: in spinor notation.

660: Finally, the additional coordinate is

661: \begin{equation} 

662: \tilde y_2=x_2- \frac{R \rho_2}{z\rho_1-\rho_2} \,.

663: \end{equation} 

664: 

665: Note that the first term is the instanton in the {\em singular} gauge,

666: the second is the anti-instanton in the {\em regular} gauge, and the third 

667: is a ``correction'' term.

668: The benefit of this representation is that the same 't Hooft symbol appears 

669: in all three terms, and the entire construction originates from conformal 

670: transformation of a spherically symmetric configuration in which $\bar I,I$ 

671: share the same center. 

672: An unfortunate feature of this expression is that time-reversal symmetry

673: is far from obvious, and it is not clear that the electric field at the 

674: mid-plane vanishes. 

675: However, this is in fact the case and the field at $t=0$ can be interpreted

676: as a turning state.

677: 

678: Although all three trial functions are similar at large

679: \IA separation $T$, they are drastically different at $T\sim \rho$.

680: The Yung ansatz is the only one which allows us to reasonably study 

681: the effects of a large change in topological number.

682: The variation of the Chern-Simons number of the turning state ($t=0$)

683: as a function of the \IA separation $T$ can be seen in

684: Fig.~\ref{NCSvsR}. 

685: In this case we scan the entire range $[0,1]$.

686: \begin{figure}[tb]

687: \begin{center}

688: \epsfig{file=fig6.eps, width=80mm}

689: \end{center}

690: \caption {\label{NCSvsR}

691: $N_{CS}$ versus the distance between $\bar I I$ centers $T$ in the Yung ansatz. 

692: }

693: \end{figure}

694: 

695: We now proceed with a more detailed study of the static turning states,

696: residing on the $t=0$ 3-plane. 

697: The simplest observable is the shape of the corresponding

698: magnetic field squared, or the energy density distribution, shown 

699: in Fig.~\ref{BB_profile} for few selected values of $\bar I I $

700: distance $T$. 

701: Note that the curve for $T=2$ (the most like the sphaleron)

702: show indeed the largest magnitude of the magnetic field.

703: The shape is however rather uniform. 

704: Note also that, unlike the case of the faulty sum and ratio trial 

705: functions, for smaller $T$ the field strength decreases, 

706: ultimately disappearing at $T=0$.  

707: \begin{figure}

708: \begin{center}

709: \epsfig{file=fig7.eps, width=80mm}

710: \end{center}

711: \caption {\label{BB_profile}

712: The ${\cal B}(r)^2$ profile, not normalized, 

713: for the four values of the $\bar I I $

714: distance $T$ (in units of $\rho$) indicated in the legend.

715: }

716: \end{figure}

717: 

718: The energy and energy density of the turning state configurations is therefore

719: rather different for different $T$. 

720: However, as seen from Fig.~\ref{BB_profile}, the physical sizes of these

721: objects are different as well. 

722: As classic Yang-Mills theory has scale invariance, one may wish to make the

723: more natural comparison of a scale-invariant combination,

724: the energy times the r.m.s. radius, $R$, defined as

725: \begin{equation} 

726: R^2= {\int d^3r\, r^2 {\cal B}^2 \over \int d^3r\, {\cal B}^2}.

727: \end{equation} 

728: In these terms, the normalized energy is

729: \begin{equation} 

730: ER=\frac{1}{2}

731: \left[\int d^3r r^2 {\cal B}^2 \times \int d^3r {\cal B}^2\right]^{1/2}.

732: \end{equation} 

733: This quantity is plotted versus the topological charge difference in 

734: Fig.~\ref{ERvsNCS}, and indeed displays a parabolic-looking maximum 

735: near $N_{CS}=1/2$. 

736: \begin{figure}[h]

737: \begin{center}

738: \epsfig{file=fig8.eps, width=70mm}

739: \epsfig{file=fig9.eps, width=70mm}

740: \end{center}

741: \caption {\label{ERvsNCS}

742: The normalized energy, $ER$, versus the Chern-Simons number

743: for the Yung ansatz. Plot (a) shows the positions of the turning states

744: for various $T$, while (b) combines many points along the path

745: ($t\ne0$); their small spread means that Yung ansatz is nearly

746: going directly uphill, thus passing via the same points for different $T$.  

747: }

748: \end{figure}

749: 

750: Instead of only looking at the static $t=0$ (and zero electric field)

751: turning states, one can instead follow the (scale invariant) energy $ER$ 

752: and the  Chern-Simons number as a function of time $t$ along each each path. 

753: As expected, {\em all} the paths in Fig.~\ref{ERvsNCS}(b), for any $T$, 

754: actually climb nearly exactly the same cliff, as they propagate into larger

755: values of our topological coordinate.

756: 

757: \section{Turning States from Constrained Minimization}

758: 

759: We will now define turning states in terms of the gauge field,

760: which connect the Euclidean and Minkowski domains of

761: the field's path. 

762: The turning state is characterized by the

763: condition that the generalized momentum, which in the

764: ${\cal A}_0=0$ gauge coincides with the chromoelectric field,

765: vanishes or, equivalently, that all first time derivatives of the

766: spatial field components are zero. 

767: Using the notation introduced in Section IB,

768: this in turn means that

769: $\partial_0\phi=\partial_0\alpha=\partial_0 A_1=0$ at the

770: time when the transition occurs. 

771: From now on we assume that moment to be $t=0$.

772: 

773: At any given time it is possible to use the special

774: gauge transformation, Eq.~(\ref{gauge_transform}), with a time-independent

775: angle $\beta$ to gauge out $A_1(r)$, still within the $A_0=0$ gauge.

776: At $t=0$ the energy of the field can thus be written

777: \begin{equation} 

778: \label{Estat}

779: E=\frac{4\pi}{g^2}\int dr \left[(\partial_r\phi)^2+\phi^2(\partial_r\alpha)^2+

780: \frac{(1-\phi^2)^2}{2r^2}\right] .

781: \end{equation} 

782: 

783: We now address the question of the {\em minimal}

784: potential energy of static Yang-Mills field, consistent with

785: the appropriate constraints:

786: (i) a fixed value of the (corrected) Chern-Simons number, and

787: (ii) a given value of the r.m.s. size.

788: The former parametrizes the position of configuration on the topological

789: scale, and the latter is needed to break dilatation symmetry of the problem,

790: which otherwise prevents any configuration of finite size from being the 

791: minimum of the energy.

792: 

793: We will break the scale invariance of the theory by setting a

794: requirement that the ratio

795: \begin{equation}

796: \langle r^2 \rangle =\frac{\int d^3x\, r^2 { B}^2}{\int d^3x\, { B}^2}

797: \end{equation}

798: has a particular value, $\rho^2$, for the static solution we seek.

799: To keep both the Chern-Simons number and mean radius constant we introduce

800: Lagrange multipliers and search for their minimal combination of\footnote{

801: Without the term introduced to fix the Chern-Simons number

802: or, equivalently, for zero corresponding Lagrange multiplier,

803: this problem would be equivalent to the SU(2) sphaleron on a 3-d sphere 

804: that was solved by Smilga \cite{Smilga}.}

805: \begin{eqnarray}

806: \tilde{E}&=&\frac{4\pi}{g^2}

807:          \! \int \! dr \left(1+\frac{r^2}{\rho^2}\right)

808:                 \!  \left[

809:                           (\partial_r \phi)^2

810:                         + \phi^2 (\partial_r\alpha)^2

811:                         + \frac{(1-\phi^2)^2}{2r^2}

812:                   \right]\nonumber\\

813: &&+\frac{\eta}{2\pi}\int dr (1-\phi^2)\partial_r\alpha \,,

814: \end{eqnarray}

815: where the tilde denotes the constrained energy.

816: It is convenient to introduce a dimensionless variable, 

817: \[

818: \xi=2\arctan\left(\frac{r}{\rho}\right)-\frac{\pi}{2} \,,

819: \]

820: so that

821: \begin{eqnarray}

822: \tilde{E}=\frac{8\pi}{g^2}

823:           \int\limits_{-\pi/2}^{\pi/2} d\xi

824:                  &\Big[&

825:                           (\partial_\xi \phi)^2

826:                         + \phi^2 (\partial_\xi\alpha)^2

827:                         + \frac{(1-\phi^2)^2}{2\cos^2\xi}\nonumber\\

828:       &&+\kappa(1-\phi^2)\partial_\xi\alpha\Big]

829: \end{eqnarray}

830: where $\kappa=\eta\rho g^2/(32\pi^2)$.

831: 

832: The Euler-Lagrange equations for the remaining fields are

833: \begin{equation} 

834: \label{EL1}

835: \partial^2_\xi\phi-\phi(\partial_\xi\alpha)^2+

836: \frac{(1-\phi)^2\phi}{\cos^2\xi}+2\kappa\phi\partial_\xi\alpha=0

837: \end{equation} 

838: and

839: \begin{equation} 

840: \label{EL2}

841: \partial_\xi(\phi^2\partial_\xi\alpha)+\kappa\partial_\xi(1-\phi^2)=0 \,.

842: \end{equation} 

843: Finiteness of the energy requires the boundary conditions

844: $\phi^2(-\pi/2)=\phi^2(\pi/2)=1$.

845: Eq.~(\ref{EL2}) integrates to

846: \begin{equation} 

847: \partial_\xi\alpha=-\kappa\frac{1-\phi^2}{\phi^2} \,,

848: \end{equation} 

849: with a vanishing integration constant as follows from the form of

850: the energy. 

851: After the substitution of $\partial_\xi\alpha$ into Eq.~(\ref{EL1}) one has

852: \begin{equation} 

853: \partial^2_\xi\phi+\frac{(1-\phi^2)\phi}{\cos^2\xi}=

854: \kappa^2\frac{1-\phi^4}{\phi^3} \,.

855: \end{equation} 

856: A solution to this equation exists for $-1<\kappa<1$,

857: \begin{equation} 

858: \phi^2=1-(1-\kappa^2)\cos^2\xi  \,.

859: \end{equation} 

860: Hereafter we assume that $\phi$ is positive.

861: 

862: In term of the usual $r$ coordinate, we have instead

863: \begin{eqnarray}

864: \phi(r) &=& \left(1-(1-\kappa^2)\frac{4\rho^2 r^2}{(r^2+\rho^2)^2}\right)^{1/2}

865: \nonumber\\ 

866: \partial_r\alpha(r)&=&-2\kappa\frac{1-\phi^2}{\phi^2}\frac{\rho}{r^2+\rho^2}\,.

867: \label{solved}

868: \end{eqnarray}

869: The sphaleron solution corresponds to $\kappa=0$ and

870: \begin{equation}

871: \phi(r)=\frac{|r^2-\rho^2|}{r^2+\rho^2}\,,\quad \alpha(r)=\pi\theta(r-\rho).

872: \end{equation}

873: 

874: For any $\kappa$ and mean squared radius

875: $\langle r^2 \rangle=\rho^2$, the potential energy density is

876: \begin{equation}

877: \frac{1}{2} B^2 = 24 \rho^4 \frac{(1-\kappa^2)^2}{(r^2+\rho^2)^4} \,,

878: \end{equation}

879: the integral of which is the potential (magnetic) energy of the 

880: static configuration,

881: \begin{equation}

882: E_B = 3\pi^2 \frac{(1-\kappa^2)^2}{g^2\rho} \,. 

883: \end{equation}

884: The corrected Chern-Simons number, computed from the first term of 

885: Eq.~(\ref{csdef}), is

886: \begin{equation}

887: \tilde{N}_{CS}= \frac{1}{4}{\rm sign}(\kappa)(2+|\kappa|)(1-|\kappa|)^2\,.

888: \end{equation}

889: Figure~\ref{dima's_profile} shows the profile of the potential energy

890: $E_{B}$ versus $\tilde{N}_{CS}$. 

891: It is very similar, although not identical, to the findings of 

892: the preceding section (see Fig.~\ref{ERvsNCS}) 

893: where Yung's ansatz was used for forced paths.

894: \begin{figure}

895: \begin{center}

896: \epsfig{file=fig10.eps, width=50mm}

897: \end{center}

898: \caption {\label{dima's_profile}

899: The potential energy

900: $E_{B}$ versus $\tilde{N}_{CS}$, for the analytic turning state

901: solution of Eq.~(\ref{solved}).

902: }

903: \end{figure}

904: 

905: \section{Explosions of the Turning States: Analytic Treatment}

906: 

907: We are now going to use the static field configuration,

908: found in previous section, as an initial condition for

909: real-time, Minkowski evolution of the gauge field. 

910: Let us first consider the equations of motion in the 1+1 dimensional

911: dynamical system.

912: Variation of the action, Eq.~(\ref{action}), gives

913: \begin{eqnarray}

914: \label{ELt1}

915: \partial_\mu\partial^\mu\phi+\phi(\partial_\mu\alpha-A_\mu)^2

916: +\frac{(1-\phi^2)\phi}{r^2}&=&0

917: \\

918: \label{ELt2}

919: \partial^\mu\left[\phi^2\left(\partial_\mu\alpha-A_\mu\right)\right] &=& 0

920: \\

921: \phi^2(\partial_1\alpha-A_1)-

922: \partial_0\left[\frac{r^2}{2}(\partial_0A_1-\partial_1A_0)\right]&=&0

923: \nonumber\\

924: \label{ELt3}

925: \phi^2(\partial_0\alpha-A_0)-

926: \partial_1\left[\frac{r^2}{2}(\partial_0A_1-\partial_1A_0)\right]&=&0.

927: \end{eqnarray}

928: The solution of Eq.~(\ref{ELt2}) has the form

929: \begin{eqnarray}

930: \phi^2(\partial_0\alpha-A_0) &=& -\partial_1\psi \nonumber\\

931: \phi^2(\partial_1\alpha-A_1) &=& -\partial_0\psi \,,

932: \end{eqnarray}

933: where $\psi(r,t)$ is an arbitrary smooth function.

934: Eqs.~(\ref{ELt3}) are consistent with this solution if

935: \begin{equation} 

936: \partial_0A_1-\partial_1A_0=-\frac{2\psi}{r^2}

937: \end{equation} 

938: 

939: Now, combining Eq.~(\ref{ELt2}) and Eqs.~(\ref{ELt3}) one has

940: \begin{equation} 

941: \label{psi_rt}

942: \partial^\mu\left(\frac{\partial_\mu\psi}{\phi^2}\right)=

943: \partial_0A_1-\partial_1A_0=\frac{2\psi}{r^2},

944: \end{equation} 

945: which can be viewed as a necessary and sufficient condition

946: for $\psi$ to be a solution for Eq.~(\ref{ELt2}) and Eqs.~(\ref{ELt3})

947: simultaneously.

948: Eq.~(\ref{ELt1}) is now

949: \begin{equation} 

950: \label{phi_rt}

951: \partial_\mu\partial^\mu\phi-\frac{(\partial_\mu\psi)^2}{\phi^3}

952: +\frac{(1-\phi^2)\phi}{r^2}=0\,.

953: \end{equation} 

954: 

955: The initial conditions for Eqs.~(\ref{psi_rt}) and (\ref{phi_rt}) are

956: \begin{eqnarray}

957: \phi(r,0)&=&\phi(r)\,,

958: \nonumber\\

959: \partial_0\phi(r,t)|_{t=0}&=&0\,,

960: \nonumber\\

961: \partial_1\psi(r,0)&=&-\phi(r)^2\partial_0\alpha(r)=0 

962: \Rightarrow \psi(r,0)=0,

963: \nonumber\\

964: \partial_0\psi(r,t)|_{t=0}&=&-\phi(r)^2\partial_1\alpha(r),\nonumber

965: \end{eqnarray}

966: where the $t$-independent fields on the right sides of the equations

967: are the static solutions of $\phi$ and $\alpha$ from the previous section.

968: 

969: As with static solutions, it is more convenient to discuss the time-evolution

970: equations in hyperbolic coordinates. 

971: Let us choose $\omega$ and $\tau$ such that

972: \begin{equation}

973: \label{tau-omega}

974: r=\frac{\rho\cos\omega}{\cos\tau-\sin\omega} \,, 

975: \quad t=\frac{\rho\sin\tau}{\cos\tau-\sin\omega} \,.

976: \end{equation}

977: The physical domain of $0<r<\infty$ and $-\infty<t<\infty$ 

978: is covered by $-\pi/2<\omega<\pi/2$ and $-\pi/2+\omega<\tau<\pi/2-\omega$.

979: For $t>0$, the corresponding domain is

980: $-\pi/2<\omega<\pi/2$ and $0<\tau<\pi/2-\omega$. 

981: This change of variables (\ref{tau-omega}) is a conformal one.

982: 

983: In the new variables Eqs.~(\ref{psi_rt}) and (\ref{phi_rt}) become\footnote{

984: One can find a discussion of Eqs.~(\ref{eqs_ot}) and some of

985: its solutions in \cite{FKS}.}

986: \begin{eqnarray}

987: \label{eqs_ot}

988: -\pt^2\phi+\po^2\phi-\frac{(\pt\psi)^2-(\po\psi)^2}{\phi^3}

989: +\frac{(1-\phi^2)\phi}{\cos^2\omega} &=& 0 

990: \nonumber\\

991: -\pt\frac{\pt\psi}{\phi^2}+\po\frac{\po\psi}{\phi^2}-

992: \frac{2\psi}{\cos^2\omega} &=& 0 \,.

993: \end{eqnarray}

994: 

995: Before solving these equations let us note that it is possible to predict

996: the large-$t$ behavior of gauge field from the form of 

997: the conformal transformation (\ref{tau-omega}). 

998: Indeed, the $t\rightarrow\infty$ limit corresponds to the line

999: $\tau=\pi/2-\omega$ on the $(\omega,\tau)$ plane. 

1000: If one now takes the limit $|r-t|\rightarrow\infty$ 

1001: (regardless of the limit for $|r-t|/t$),

1002: the position on $(\omega,\tau)$ plane is either 

1003: $\omega\rightarrow -\pi/2\,,\,\tau\rightarrow 0$ or $\omega\rightarrow \pi/2

1004: \,,\, \tau\rightarrow\pi$. 

1005: This means that the entire line $\tau=\pi/2-\omega$

1006: corresponds to space-time points with finite differences between

1007: $r$ and $t$ and, therefore, if $\phi$ and $\psi$ are smooth functions

1008: of $\omega$ and $\tau$, then for asymptotic times the field is concentrated

1009: near the $r=t$ line.

1010: This corresponds to the fields expanding as a thin shell in space.

1011: 

1012: We must now supply Eqs.~(\ref{eqs_ot}) with initial conditions, which are

1013: \begin{eqnarray}

1014: \label{init}

1015: \phi(\omega,\tau=0)^2 &=& 1-(1-\kappa^2)\cos^2\omega

1016: \nonumber\\

1017: \pt\phi(\omega,\tau)|_{\tau=0} &=& 0

1018: \nonumber\\

1019: \psi(\omega,\tau=0) &=& 0

1020: \nonumber\\

1021: \pt\psi(\omega,\tau)|_{\tau=0} &=& \frac{\rho}{1-\sin\omega}

1022: \partial_t\psi(\omega,\tau)|_{t=0} 

1023: \nonumber\\

1024: &=&\kappa(1-\kappa^2)\cos^2\omega \,.

1025: \end{eqnarray}

1026: One of the solutions of Eqs.~(\ref{eqs_ot}), first found in 1977 by

1027: L\"uscher \cite{Luscher} and Schechter \cite{Schechter}, is

1028: \begin{eqnarray}

1029: \phi(\omega,\tau)^2 &=& 1-(1-q^2(\tau))\cos^2\omega \nonumber\\

1030: \psi(\omega,\tau) &=& \frac{\dot{q}(\tau)}{2}\cos^2\omega \,,

1031: \end{eqnarray}

1032: with a function $q(\tau)$ that satisfies

1033: \begin{equation}

1034: \label{ddotq}

1035: \ddot{q}-2q(1-q^2)=0 \,.

1036: \end{equation}

1037: This is the equation for a one-dimensional particle

1038: moving in double-well potential of the form $U(q)=(1-q^2)^2/2$.

1039: 

1040: We now have to check that the L\"uscher-Schechter solution

1041: satisfies the initial conditions, (\ref{init}).

1042: This is indeed the case if one identifies $q(0)=\kappa$ 

1043: and takes $\dot{q}(0)=0$.

1044: For the initial condition of this type

1045: ({\em i.e.} for energy $\varepsilon=\dot{q}^2/2+U(q)<1/2$),

1046: the solution of Eq.~(\ref{ddotq}) is

1047: \begin{equation}

1048: \label{solution}

1049: q(\tau)=\tilde{q}\dn\left(\tilde{q}(\tau-\tau_0), k\right),

1050: \end{equation}

1051: where $\dn$ is Jacobi's function and

1052: $\tilde{q}=\sqrt{2-\kappa^2}$ is the second stopping point

1053: for a particle in the potential $U(q)$.

1054: We have also defined

1055: \[

1056: k^2=2\frac{1-\kappa^2}{2-\kappa^2} \quad {\rm and}\quad

1057: \tau_0\tilde{q}=\frac{T}{2} \,,

1058: \]

1059: where $T$, the period of oscillations in the potential $U(q)$, is 

1060: $T=2K(k)$, with $K(k)$ being the complete elliptic integral of the first kind.

1061: The idea is, of course, that ``oscillations'' in $\tau$ begin

1062: from the rest point, close to $\tau = 0$.

1063: 

1064: Let us now look at several properties of the solution for large times.

1065: The solution (\ref{solution}) is apparently regular in the $(\omega, \tau)$

1066: plane, and therefore for large times the field is concentrated near $r=t$. 

1067: At asymptotic times the energy density, $e(r,t)$, is given by

1068: \begin{equation}

1069: 4\pi e(r,t) = \frac{8\pi}{g^2\rho^2}(1-\kappa^2)^2

1070: \left(\frac{\rho^2}{\rho^2+(r-t)^2}\right)^3 \,.

1071: \end{equation}

1072: The change in topological charge is

1073: \begin{eqnarray}

1074: \Delta Q &=& \int\limits_0^{\infty} d^3x dt \,\partial_\mu K^\mu\nonumber\\

1075: &=& \frac{1}{2\pi}\int dr dt

1076: \left[-\partial^2_t\psi+\partial^2_r\psi-\frac{2\psi}{r^2}\right]\nonumber\\

1077: &=& \frac{\pi}{2}\kappa(3-\kappa^2)

1078: -{\rm sign}(\kappa)\arccos\left(

1079: \frac{\mbox{\rm cn}(\tilde{q}\pi, k)}{\dn(\tilde{q}\pi, k)}\right) .

1080: \end{eqnarray}

1081: The evolution of $\tilde{N}_{CS}$ begins from time $t=0$, where

1082: \begin{equation}

1083: \tilde{N}_{CS}(0)=\frac{1}{4}{\rm sign}(\kappa)(1-|\kappa|)^2(2+|\kappa|) \,,

1084: \end{equation}

1085: and as $t\rightarrow\infty$ its limit is

1086: $\tilde{N}_{CS}(\infty)=\tilde{N}_{CS}(0)+\Delta Q$.

1087: 

1088: We now estimate number of gluons produced by the described evolution. 

1089: In $\phi,\psi$ language the chromoelectric and chromomagnetic

1090: fields are

1091: \begin{eqnarray}

1092: { E}^a_j  &=&

1093: \frac{1}{r} \left(\partial_t\phi\sin\alpha -

1094: \frac{\partial_r\psi\cos\alpha}{\phi}\right)\Theta^a_j \nonumber\\

1095: &&+ \frac{1}{r} \left(\partial_t\phi\cos\alpha +

1096: \frac{\partial_r\psi\sin\alpha}{\phi}\right)\Pi^a_j +

1097: \frac{2\psi}{r^2}\Sigma^a_j \,,

1098: \label{electric2}

1099: \end{eqnarray}

1100: \begin{eqnarray}

1101: { B}^a_j &=&

1102: -\frac{1}{r} \left(\partial_r\phi\cos\alpha +

1103: \frac{\partial_t\psi\sin\alpha}{\phi}\right)\Theta^a_j \nonumber \\

1104: && + \frac{1}{r}\left(\partial_r\phi\sin\alpha -

1105: \frac{\partial_t\psi\cos\alpha}{\phi}\right)\Pi^a_j +

1106: \frac{1-\phi^2}{r^2}\Sigma^a_j \,.

1107: \label{magnetic2}

1108: \end{eqnarray}

1109: Terms proportional to $\Sigma^a_j$ are longitudinal and die out as 

1110: $t\rightarrow\infty$. 

1111: The remainder is a purely transverse field.

1112: Now let us take into account that for $t\rightarrow\infty$,

1113: $\partial_r\phi\rightarrow\partial_t\phi$, and the same for $\psi$.

1114: Therefore

1115: \begin{eqnarray}

1116: { E}^a_j &\rightarrow&

1117: \frac{1}{r} \left(\partial_r\phi\sin\alpha -

1118: \frac{\partial_r\psi\cos\alpha}{\phi}\right)\Theta^a_j  \nonumber \\

1119: &&+   \frac{1}{r} \left(\partial_r\phi\cos\alpha +

1120: \frac{\partial_r\psi\sin\alpha}{\phi}\right)\Pi^a_j \,,

1121: \label{electric3}

1122: \end{eqnarray}

1123: \begin{eqnarray}

1124: { B}^a_j &\rightarrow&

1125: -\frac{1}{r} \left(\partial_r\phi\cos\alpha +

1126: \frac{\partial_r\psi\sin\alpha}{\phi}\right)\Theta^a_j  \nonumber \\

1127: &&+   \frac{1}{r} \left(\partial_r\phi\sin\alpha -

1128: \frac{\partial_r\psi\cos\alpha}{\phi}\right)\Pi^a_j \,.

1129: \label{magnetic3}

1130: \end{eqnarray}

1131: The main result becomes apparent when we choose a gauge where

1132: \[\phi\partial_r\phi\cos\alpha + \partial_r\psi\sin\alpha=0\,, \]

1133: in which 

1134: \begin{eqnarray}

1135: { E}^a_j &\rightarrow&

1136: \frac{1}{r}

1137: \sqrt{\frac{(\partial_r\psi)^2}{\phi^2}+(\partial_r\phi)^2}\Theta^a_j 

1138: \nonumber\\ 

1139: &\rightarrow&

1140: \frac{1-\kappa^2}{r\rho}\left(\frac{\rho^2}{\rho^2+(r-t)^2}\right)^{3/2}

1141: \Theta^a_j \,,

1142: \label{electric4}

1143: \end{eqnarray}

1144: \begin{equation}

1145: { B}^a_j \rightarrow

1146: \frac{1-\kappa^2}{r\rho}\left(\frac{\rho^2}{\rho^2+(r-t)^2}\right)^{3/2}

1147: \Pi^a_j \,.

1148: \label{magnetic4}

1149: \end{equation}

1150: 

1151: We now perform a fourier transform, finding

1152: \begin{eqnarray}

1153: { E}^a_j(\vec{k})

1154: &=& 4\pi\rho(1-\kappa^2)K_1(\omega\rho)\Theta^a_j\nonumber\\

1155: { B}^a_j(\vec{k})

1156: &=& 4\pi\rho(1-\kappa^2)K_1(\omega\rho)\Pi^a_j\,,

1157: \end{eqnarray}

1158: where $\Theta^a_j$ and $\Pi^a_j$ are the color/space 

1159: projectors in momentum space analogous to those in coordinate space

1160: (\ref{projectionops}), the frequency $\omega = |\vec{k}|$,

1161: and $K_1$ is a Bessel function.

1162: One can easily verify that ${ B}^a_j=\epsilon_{jlm}k_l{ E}^a_m/k$, as 

1163: is required for a radiation field. 

1164: 

1165: \section{Explosions of the Turning States:  Numerical Studies}

1166: 

1167: In the previous section we studied the asymptotic behavior of 

1168: turning states constrained by sphaleron size and Chern-Simons number.

1169: In this section we consider step-by-step evolution of the turning

1170: states, a numerical analysis similar to sphaleron decay in electroweak

1171: theory \cite{HK,Zad}.

1172: The numerical approach naturally allows for mathematical flexibility,

1173: and it is used to consider the decay of static states which replace the

1174: unphysical power-law behavior of the fields at large distance with

1175: a phenomenologically more appropriate exponential tail.

1176: The classical field configurations are thus fixed in size indirectly by a 

1177: mass parameter, a constraint which is subsequentially

1178: relaxed as the state quickly decays into free-streaming gluons.

1179: 

1180: As a result of scale invariance, the QCD instanton is of indeterminate

1181: size.

1182: While it is clear from phenomenology and lattice studies that the instanton

1183: vacuum favors a somewhat narrow size distribution centered at $\bar\rho

1184: \simeq 0.3$ fm, the reason for this is yet unknown, although

1185: it is presumably due to interactions between instantons.

1186: It is thus natural that related classical objects, born in some way from

1187: the excitation of instantons, share a similar size.

1188: We arrange this by introducing a phenomenological gluon mass term in 

1189: the initial configuration which is promptly relaxed as this unstable

1190: configuration begins to decay.

1191: We stress that the relaxation of this size constraint does not initiate

1192: the explosion; we will show that the turning state is an unstable 

1193: configuration regardless of the mass term's presence.

1194: 

1195: The similarity between this procedure and electroweak sphaleron decay is 

1196: clear, but not mathematically continuous.

1197: For the Higgs mechanism, the sphaleron size is constrained by the scalar

1198: vacuum condensate which introduces an effective mass for the gauge fields.

1199: This condensate vanishes at the sphaleron center, where the classical

1200: gauge field action is maximal.

1201: This feature persists in the limit of infinite gauge-Higgs coupling, 

1202: when the Higgs field is fixed at its VEV constant for all other points in space.

1203: Inserting a mass term for the gauge fields by hand, as we do here for QCD,

1204: is thus very similar to this infinite coupling limit of electroweak dynamics,

1205: the only difference being that the mass is finite in {\it all} of space.

1206: This leads to a difference in field behavior at the origin.

1207: 

1208: We begin with the Yang-Mills action, with a phenomenological mass term 

1209: added, and will look for static solutions with

1210: spherical symmetry in Minkowski space.

1211: The action is written:

1212: \begin{eqnarray}

1213: S &=& \frac{4\pi}{g^2}\int dt dr \Bigg\{ \dot{\phi}_1^2 + \dot{\phi}_2^2 

1214: + \frac{1}{2} r^2 \dot{A_1}^2 - (\phi_1')^2 - (\phi_2')^2 \nonumber\\

1215: &&-\frac{\left(1-\phi_1^2 - \phi_2^2\right)^2}{2 r^2} 

1216: - 2 A_1\left(\phi_1\phi_2' - \phi_1\phi_2'\right) 

1217: \label{gcaction} \\ &&

1218: - A_1^2\left( \phi_1^2 + \phi_2^2 \right) 

1219: - m^2 \left[ \left(1+\phi_1\right)^2 + 

1220: \phi_2^2 + \frac{1}{2} r^2 A_1^2 \right] \Bigg\}. \nonumber

1221: \end{eqnarray}

1222: As before, we have taken a spherical ansatz similar to Witten's 

1223: \cite{Witten_ans} for the gauge field.

1224: Following the notations of Eq.~(\ref{sph_ansatz}),

1225: \begin{equation}

1226: A = \frac{1-\phi_1}{r} \,, \quad

1227: B = \frac{\phi_2}{r}  \,, \quad

1228: C = A_1 \,,  \quad

1229: D = A_0  \,,

1230: \end{equation}

1231: %A^{ia}(r) &=& \frac{1}{g}\left[ \frac{1-f_A}{r} \epsilon^{ija} \hat{x}^j

1232: % + \frac{f_B}{r}\left( \delta^{ia}-\hat{x}^i\hat{x}^j \right)

1233: % + f_C \hat{x}^i\hat{x}^a \right]

1234: % \nonumber\\

1235: %A_0^a(r) &=& \frac{1}{g} G \hat{x}^a \,,

1236: where all scalar functions depend on both $r$ and $t$.

1237: We initially work in the temporal gauge, where $A_0(r,t) = 0$.

1238: 

1239: The equations of motion are easily obtained, and we find

1240: \begin{eqnarray}

1241: && \ddot{\phi_1} - \phi_1'' + \frac{\phi_1}{r^2}\left(\phi_1^2+\phi_2^2-1\right)

1242: + \phi_1 A_1^2 - 2 \phi_2' A_1 \nonumber\\

1243: && \qquad - \phi_2 A_1' + m^2\left(\phi_1 + 1\right) = 0 \,,

1244: \nonumber \\

1245: && \ddot{\phi_2} - \phi_2'' + \frac{\phi_2}{r^2}\left(\phi_1^2+\phi_2^2-1\right)

1246: + \phi_2 A_1^2 + 2 \phi_1' A_1 \nonumber\\

1247: && \qquad + \phi_1 A_1' + m^2 \phi_2 = 0 \,,

1248: \nonumber \\

1249: && \ddot{A_1} + \frac{2}{r^2} A_1 \left( \phi_1^2 + \phi_2^2 \right)

1250: + \frac{2}{r^2} \left( \phi_1' \phi_2 - \phi_1\phi_2' \right)

1251: \nonumber\\ && \qquad 

1252: + m^2 A_1  = 0 \,.

1253: \label{eom}

1254: \end{eqnarray}

1255: 

1256: In the static limit, we seek a purely magnetic turning state solution.

1257: One can be found for the $\phi_1$ field component from the equation

1258: \begin{equation}

1259: \phi_1'' - \frac{\phi_1}{x^2}\left( \phi_1^2 - 1\right) 

1260: - \left(\phi_1+1\right) = 0 \,,

1261: \end{equation}

1262: which is simply Eqs.~(\ref{eom}) with $\phi_2 = A_1 = 0$ and written

1263: in terms of the dimensionless variable $x \equiv mr$.

1264: With the boundary conditions

1265: \begin{eqnarray}

1266: &\phi_1(x) \rightarrow 1 \:& {\rm as} \: x\rightarrow 0 \,, \nonumber\\

1267: &\phi_1(x) \rightarrow -1 \:& {\rm as} \: x\rightarrow \infty \,, \nonumber

1268: \end{eqnarray}

1269: A numerical solution is easily obtained and shown in Fig.~\ref{fafig}.

1270: This is quite similar to the approximate electroweak solution found by 

1271: Klinkhammer and Manton \cite{KM} in the limit of infinite Higgs self-coupling.

1272: The primary difference is the behavior near the origin, which in the QCD

1273: case involves a logarithm for $x \ll 1$:

1274: \begin{equation}

1275: \phi_1(x) = 1 + \frac{2}{3} x^2 \ln x - \alpha x^2 + {\cal O}

1276: \left( x^4 \ln x\right) \,,

1277: \end{equation}

1278: where $\alpha = 1.98$ was determined numerically.

1279: Defining the turning state's size as the radius at which the profile is at

1280: half its maximum ({\it i.e.} where it crosses the origin), we find

1281: $m\rho \simeq 0.9$.  

1282: We match this to the size of the average instanton and find

1283: \begin{equation}

1284: m = 0.9 \rho^{-1} \simeq 540 \,{\rm MeV}.

1285: \end{equation}

1286: \begin{figure}[b]

1287: \begin{center}

1288: \epsfig{file=fig11.eps, width=80mm}

1289: \end{center}

1290: \caption{

1291: The static turning state solution, $\phi_1(r,0)$.

1292: }\label{fafig}

1293: \end{figure}

1294: 

1295: Before we discuss the decay of this configuration we must find the 

1296: unstable modes orthogonal to it which determine the ``downhill'' directions 

1297: in field space.

1298: This is done by solving eigenvalue equations for fluctuations in

1299: the fields $\phi_2$ and $A_1$ in the presence of the turning state 

1300: configuration.

1301: 

1302: We take the terms linear in $\phi_2$ and $A_1$ from Eqs.~(\ref{eom}) and 

1303: require

1304: \begin{equation}

1305: \ddot{\phi_2}(x,t) = - \omega^2 \phi_2(x,t) \,,\quad

1306: \ddot{A_1}(x,t) = - \omega^2 A_1(x,t)\,.

1307: \end{equation}

1308: We then have the eigenvalue equations:

1309: \begin{eqnarray}

1310: \phi_2'' + \left(\Omega^2 - 1 + \frac{1}{x^2}\right) \phi_2

1311: + 2 \phi_1' A_1 + \phi_1 A_1'  &=& 0 \,,\nonumber\\

1312: \left( \Omega^2 - 1 - \frac{2}{x^2}\phi_1^2 \right) A_1

1313: - \frac{2}{x^2} \left( \phi_1' \phi_2 - \phi_1\phi_2' \right) &=& 0 \,,

1314: \label{unst_eom}

1315: \end{eqnarray}

1316: where $\phi_1$ is the classical solution in Fig.~\ref{fafig} and the 

1317: dimensionless frequency is $\Omega = \omega/m$.

1318: The longitudinal field component may be eliminated with

1319: \begin{equation}

1320: A_1 = \frac{2\left( \phi_1' \phi_2 - \phi_1 \phi_2'\right)}

1321: {\left(\Omega^2-1\right)x^2 - 2 \phi_1^2}\,.

1322: \end{equation}

1323: Substituting this into the first of Eqs.~(\ref{unst_eom}), we find

1324: the behavior near the origin:

1325: \begin{equation}

1326: \phi_2(x) = c x^{ \frac{1}{2}\left(1-\sqrt{1-8y}\right)} \,,

1327: \end{equation}

1328: where 

1329: \[

1330: y = \frac{1+\Omega^2}{1-\Omega^2}

1331: \]

1332: and $c$ is an arbitrary normalization.

1333: Both fields vanish as $x\rightarrow\infty$.

1334: 

1335: We have solved these equations numerically, finding

1336: the wave functions plotted in Fig.~\ref{unstfig} with the frequency

1337: \begin{equation}

1338: \omega^2 = - 3.4 m^2 \,.

1339: \end{equation}

1340: The function $A_1$ is logarithmically divergent at the origin, reflecting

1341: the difference between this massive model and the electroweak case

1342: \cite{AKY2} in 

1343: which the Higgs always vanishes at the origin\footnote{The dominant unstable

1344: mode in the electroweak as $\lambda \rightarrow \infty$ is the 

1345: scalar field component orthogonal to the condensate \cite{AKY2}.  

1346: This mode is absent here.}.

1347: There is therefore no smooth continuation between this model and the 

1348: electroweak in the limit of large coupling.

1349: \begin{figure}[bt]

1350: \begin{center}

1351: \epsfig{file=fig12.eps, width=80mm}

1352: \end{center}

1353: \caption{

1354: The unstable eigenmodes, $\phi_2(r)$ and $A_1(r)$, arbitrarily normalized.

1355: }\label{unstfig}

1356: \end{figure}

1357: 

1358: These solutions for the unstable modes, which along with the classical

1359: $\phi_1(r)$ complete our initial conditions, were put on a lattice with spacing

1360: $\Delta x = 0.01$ and evolved at time steps of $\Delta \tau = 5\times 10^{-4}$,

1361: where $\tau \equiv mt$.

1362: The unstable modes, acting as a small push to properly initiate the decay,

1363: were normalized as

1364: \begin{equation}

1365: \int dx \left( \phi_2(x)^2 + 2 x^2 A_1(x)^2 \right) = 5\times 10^{-3} \,.

1366: \end{equation}

1367: Coincident with this push we set $m=0$, in effect turning off the mass 

1368: term, since here we are interested in the turning state decaying into 

1369: the vacuum where no such term is motivated.

1370: Although this effectively removes the size constraint on

1371: $\phi_1(r,0)$, the subsequent dynamical expansion is a result of the 

1372: push in the unstable directions of this saddle-point solution rather than

1373: an inflation of the classical field configuration.

1374: 

1375: Once the real-time solutions to Eqs.~(\ref{eom}) with $m=0$

1376: have been found at a given time step, the total energy is readily computed as

1377: \begin{eqnarray}

1378: E = \frac{4\pi}{g^2}\int dr \Bigg[&& \dot{\phi}_1^2 + \dot{\phi}_2^2 

1379: + \frac{1}{2} r^2 \dot{A_1}^2 + (\phi_1')^2 + (\phi_2')^2 \nonumber\\

1380: &&+\frac{\left(1-\phi_1^2 - \phi_2^2\right)^2}{2 r^2} 

1381: - 2 A_1\left(\phi_1\phi_2' - \phi_1\phi_2'\right) 

1382: \nonumber\\ &&

1383: + A_1^2\left( \phi_1^2 + \phi_2^2 \right)  \Bigg]

1384: \label{gceden}

1385: \end{eqnarray}

1386: At every time step it was verified that the energy remains equal to that

1387: of the initial state.

1388: Taking the instanton vacuum value of

1389: \begin{equation}

1390: \frac{8 \pi^2}{g^2} = 12 \,,

1391: \end{equation}

1392: the total energy of the decaying object was calculated and found to be 

1393: \begin{equation}

1394: E = 4.62 \rho^{-1} \simeq 2.8 \,{\rm GeV}.

1395: \end{equation}

1396: 

1397: The Chern-Simons number was also computed at each time step.

1398: In our present gauge, this is written as

1399: \begin{equation}

1400: N_{CS} = \frac{1}{2\pi}\! \int\!\! dr \left[ \left(1-\phi_1\right)\phi_2'

1401: + \phi_1' \phi_2 - \left(1-\phi_1^2-\phi_2^2\right)A_1 \right].

1402: \label{gccsnum}

1403: \end{equation}

1404: The gauge invariance of changes in this quantity were verified

1405: numerically.

1406: 

1407: The energy and Chern-Simons densities, defined as the integrands of

1408: Eqs.~(\ref{gceden}) and (\ref{gccsnum}), are shown in Fig.~\ref{ecsfig}.

1409: The shell-like expansion is illustrated in these plots, as well as the

1410: similarity of the energy profile at all times.

1411: Note that these plots are in a 1+1 dimensional description, and the

1412: corresponding three-dimensional radial density differs by a factor of

1413: $1/r^2$.

1414: \begin{figure}[bt]

1415: \begin{center}

1416: \epsfig{file=fig13.eps, width=80mm}

1417: \end{center}

1418: \caption{

1419: Energy (solid lines) and Chern-Simons number densities (dashed lines) 

1420: for three times during the explosion, $t =$0.4, 3, and 6 fm.

1421: (Note the scale difference between the two quantities.)

1422: }\label{ecsfig}

1423: \end{figure}

1424: 

1425: It is clear from the curves in Fig.~\ref{ecsfig} that

1426: the Chern-Simons number changes during the decay.

1427: As shown in Fig.~\ref{csfig},

1428: our fields stabilize at long times with $\Delta N_{CS} \simeq 0.12$.

1429: From this we see that the turning state does not complete an

1430: instanton transition, which would require a return to

1431: a state with integral Chern-Simons number.

1432: Due to this freezing in the topology, nontrivial fermionic

1433: solutions will accompany the resulting Yang-Mills fields.

1434: These will be discussed elsewhere.

1435: 

1436: \begin{figure}[bt]

1437: \begin{center}

1438: \epsfig{file=fig14.eps, width=80mm}

1439: \end{center}

1440: \caption{

1441: Change in the Chern-Simons number.

1442: }\label{csfig}

1443: \end{figure}

1444: 

1445: The transition from a purely magnetic configuration to one with equal 

1446: electric and magnetic components is shown in Fig.~\ref{ebfig}, 

1447: for an early and late time in the evolution.

1448: The decay progresses rather rapidly; at $t = 1.4$ fm,

1449: the ratio $E(r)^2/B(r)^2 \simeq 0.95$ for all $r$.

1450: Thereafter the ratio continues to quickly approach unity.

1451: 

1452: \begin{figure}[bt]

1453: \begin{center}

1454: \epsfig{file=fig15.eps, width=80mm}

1455: \end{center}

1456: \caption{

1457: The electric (solid lines) and magnetic (dashed lines) fields at times

1458: $t = 0.4$ fm and $t=3$ fm.  Plotted are

1459: $\frac{1}{2} x^2 E(x)^2$ and 

1460: $\frac{1}{2} x^2 B(x)^2$; their sum is the energy density.

1461: }\label{ebfig}

1462: \end{figure}

1463: 

1464: In order to analyze the final state at late times, we

1465: we work in a gauge in which $B(r,t) = 0$.

1466: This requires the transformation to a new set of fields:

1467: \begin{eqnarray}

1468: \tilde{\phi}_1 &=& \phi_1 \cos\theta + \phi_2 \sin\theta \nonumber\\

1469: \tilde{\phi}_2 &=& -\phi_1 \sin\theta + \phi_2 \cos\theta \nonumber\\

1470: \tilde{A}_1 &=& A_1 - \theta' \nonumber\\

1471: \tilde{A}_0 &=& - \dot{\theta} \,,

1472: \end{eqnarray}

1473: where 

1474: \begin{equation}

1475: \theta = \arctan\left(\frac{\phi_2}{\phi_1}\right) \,.

1476: \end{equation}

1477: Promptly dropping the tildes, we write the total energy in terms of 

1478: the new fields,

1479: \begin{eqnarray}

1480: E &=& \frac{4\pi}{g^2} \int dr \Bigg[ \dot{\phi_1}^2 + \left(\phi_1'\right)^2

1481: + \frac{\left(1-\phi_1^2\right)^2}{2 r^2} \nonumber\\

1482: && + \frac{r^2}{2}\left(\dot{A_1}-A_0'\right)^2 

1483: + \phi_1^2 \left( A_1^2 + A_0^2 \right)\Bigg] \,.

1484: \label{gaugeenergy}

1485: \end{eqnarray}

1486: 

1487: At late times, the field strength is confined to a thin shell at radius

1488: $r = t$.

1489: Free, expanding field behavior is also observed numerically, such that

1490: \begin{equation}

1491: \dot{\Phi}(r,t) = - \Phi'(r,t)

1492: \end{equation}

1493: for each field $\phi_1$, $A_1$, and $A_0$.

1494: This simplifies the equations of motion for the latter two,

1495: \begin{eqnarray}

1496: \ddot{A_1} - \dot{A_0}' &=& - \frac{2}{r^2} \phi_1^2 A_1 \nonumber\\

1497: A_0'' - \dot{A_1}' &=& - \frac{2}{r}\left(A_0'-\dot{A_1}\right)

1498: + \frac{2}{r^2} \phi_1^2 A_1 \,,

1499: \end{eqnarray}

1500: in that the right hand sides of both vanish as the shell expands at large

1501: times.

1502: We can thus conclude that 

1503: \begin{equation}

1504: \dot{A_1} - A_0' = 0

1505: \label{a10}

1506: \end{equation}

1507: for $mt \gg 1$.

1508: 

1509: The condition for this gauge is

1510: \begin{equation}

1511: \dot{A_0} \phi_1 + 2 A_0\dot{\phi_1} - A_1'\phi_1 - 2 A_1 \phi_1' = 0 \,.

1512: \label{gaugecond}

1513: \end{equation}

1514: Combining this with Eq.~(\ref{a10}), we have

1515: \begin{equation}

1516: \phi_1\left( \dot{A_0}-A_1' \right) 

1517: - 2\left( \phi_1' A_1 - \dot{\phi_1} A_0\right) = 0 \,,

1518: \end{equation}

1519: and can deduce that 

1520: \begin{equation}

1521: A_0 + A_1 = 0 

1522: \label{zerooneeqn}

1523: \end{equation}

1524: at late times.

1525: 

1526: The contribution from $\phi_1$ will be of the form

1527: \begin{equation}

1528: \phi_1(r,t) = 1 + \varphi(r,t) \,,

1529: \end{equation}

1530: where $\varphi$ is an excitation above the vacuum.

1531: Using the result just obtained (\ref{zerooneeqn}), 

1532: its linearized equation of motion simplifies to 

1533: \begin{equation}

1534: \ddot{\varphi} - \varphi'' + \frac{2}{r^2} \varphi = 0 \,.

1535: \end{equation}

1536: The solution is of the form

1537: \begin{equation}

1538: \varphi(r,t) = \int dk \, \varphi(k) r \left[j_1(kr)+iy_1(kr)\right] \cos(kt)

1539: \,,

1540: \end{equation}

1541: with fourier amplitudes $\varphi(k)$ and the spherical 

1542: Bessel functions $j_1(z)$ and $y_1(z)$.

1543: 

1544: Although the majority of gluon radiation is carried in the $\phi_1$ field,

1545: physical quanta also lie in small excitations of the field 

1546: \[

1547: \psi(x,\tau) = \phi_1(x,\tau) A_0(x,\tau) \,,

1548: \]

1549: which encodes oscillations between the transverse degrees of freedom.

1550: From Eqs.~(\ref{gaugecond}) and (\ref{zerooneeqn}), we have a

1551: wave equation,

1552: \begin{equation}

1553: \ddot\psi - \psi'' = 0 \,.

1554: \end{equation}

1555: These harmonics contribute to the energy via the final term in 

1556: Eq.~(\ref{gaugeenergy}).

1557: 

1558: The total energy can then be written in momentum space as

1559: \begin{equation}

1560: E = \frac{16m}{g^2} \int dp \left[ p^2 \varphi(p)^2 

1561: + \psi(p)^2 \right] \,,

1562: \label{penergy}

1563: \end{equation}

1564: in terms of the dimensionless momentum $p = k/m$.

1565: The fourier amplitudes are computed from the solutions of the 

1566: spatial fields:

1567: \begin{eqnarray}

1568: \varphi(p)^2 &=& \left| \int dx\, px\left[j_1(px)+i y_1(px)\right]

1569: \varphi(x) \right|^2

1570: \nonumber\\

1571: \psi(p)^2 &=& \left| \int dx\, e^{ipx} \psi(x) \right|^2 \,.

1572: \end{eqnarray}

1573: Numerically, this expression for the energy is within 1\% of that of the 

1574: initial configuration for all times $t \ge 3$~fm, further demonstrating

1575: the rapid onset of free-field behavior.

1576: 

1577: \section{Production of Turning States in Heavy Ion Collisions}

1578: 

1579: \subsection{The pQCD Cutoff in Vacuum Versus Excited Matter}

1580: 

1581: Although this paper does not generally deal with phenomenological

1582: applications, we will eventually come to an estimate of the number of

1583: gluons produced in the explosion of a turning state. 

1584: The resulting divergence in this number cannot be resolved without 

1585: some explanation of the limited applicability of the classical

1586: Yang-Mills description. 

1587: This leads to the issue of the pQCD cutoff.

1588: 

1589: It is well known that in the QCD vacuum

1590: the ``semi-hard'' or ``substructure scale'' 

1591: $Q^2 \sim \, 1-2 \, {\rm GeV}^2$ is simultaneously

1592: the {\em lower} boundary of pQCD as well as the {\em upper} 

1593: boundary of low energy effective approaches such as chiral Lagrangians.

1594: Furthermore, in any discussion of instanton-induced reactions it is implicitly

1595: assumed that instantons were the primary source of that scale, and since 

1596: they are included explicitly no other nonperturbative cutoffs are needed.

1597: This, of course, is not strictly true as confining forces require the

1598: final state be comprised of hadrons, but we make the usual separation of

1599: scales and assume that final state interactions merely redistribute

1600: the wave functions without changing the total probabilities. 

1601: 

1602: We assert, however, that the pQCD cutoff is quite different in heavy 

1603: ion collisions.

1604: It has been argued over the years that excited matter

1605: might be in a Quark-Gluon Plasma (QGP) 

1606: phase of QCD (see e.g. Ref.~\cite{Shu_80}).

1607: Whether equilibrated or not, it is nevertheless qualitatively very 

1608: different from the QCD vacuum: instantons are suppressed, and

1609: there is neither confinement nor chiral symmetry breaking to set

1610: a nonperturbative cutoff.

1611:  

1612: Therefore the limits on Yang-Mills field description are entirely different, 

1613: and actually determined by much simpler phenomena.

1614: The QGP, a plasma-like phase, screens itself perturbatively \cite{Shu_80}. 

1615: A quasi-particle description becomes appropriate, 

1616: in which the quarks and gluons have finite effective masses. 

1617: In equilibrium and at high temperature these are ``thermal masses''; 

1618: the gluon has the well-known effective mass \cite{Shu_80}

1619: \begin{equation} 

1620: M^2_g=\frac{g^2T^2}{2} \left(\frac{N_c}{3}+\frac{N_f}{6}\right)

1621: \end{equation} 

1622: where $N_c$ and $N_f$ are the number of colors and flavors, respectively.

1623: Although this mass grows with temperature at high $T$, just above $T_c$

1624: it is actually {\em  smaller} than the pQCD cutoff in vacuum.

1625: Such non-monotonic behavior is confirmed by lattice thermodynamics data, 

1626: which can be well fitted with quasiparticle masses. 

1627: 

1628: Moreover, in the ``RHIC window'', $T_c<T<3T_c$, one finds the approximately 

1629: constant gluon and quark effective masses

1630: \cite{LH}

1631: \begin{equation} 

1632: \label{Meff}

1633: M_g\approx .4\,{\rm GeV}\,,\quad M_q\approx .3 \,{\rm GeV},

1634: \end{equation} 

1635: the first of which provides the cutoff of our classical treatment;

1636: at this scale the

1637: classical Yang-Mills action for gluons is to be modified by inclusion

1638: of an appropriate effective Lagrangian describing such screening effects,

1639: such as those suggested by Taylor and Wong \cite{TW}.

1640: 

1641: \subsection{Multiplicity and Spectra of Prompt Gluons}

1642: 

1643: With solutions for the fields at all times,

1644: in both the analytic and numerical treatments of the previous two sections, 

1645: we have analyzed the final states to determine the particle number.

1646: While we find a similar number of produced particles from both approaches,

1647: about four gluons produced per turning state, the energy distribution of

1648: this prompt glue is very sensitive to the classical configuration used

1649: to initiate the explosion.

1650: While the field solutions found using constrained quantization (Section III)

1651: and an effective mass (Section V) are qualitatively similar, both leading

1652: to an expanding shell of radiation, we find a substantial difference in final 

1653: state momentum distributions.

1654: This can be traced to the details of the two solutions as 

1655: $r\rightarrow\infty$, where we contrast a power-law behavior in

1656: Eq.~(\ref{solved}) with an exponential fall-off in the solution of

1657: Fig.~\ref{fafig}, as $\phi_1(r,0) \sim {\rm exp}(-mr)-1$.

1658: As mentioned above, we consider the second case to be of greater physical

1659: relevance, as the gluonic field is not massless phenomenologically 

1660: (and even vacuum instantons ought to have exponential tails).

1661: We now consider both results.

1662: 

1663: To find the number of gluons in each mode one compares the field 

1664: strength in the momentum representation to those which have energy $\omega$.

1665: In the evolution described analytically in Section IV, the

1666: occupation number is

1667: \begin{equation}

1668: \nu(\vec{k})

1669: =\frac{64\pi^2}{g^2}(1-\kappa^2)^2\frac{\rho^2}{\omega}K_1(\omega\rho)

1670: \label{occno}

1671: \end{equation}

1672: and the gluon energy distribution function is

1673: \begin{equation}

1674: n(\omega)=\frac{32}{g^2}({1-\kappa^2})^2\omega\rho^2K_1^2(\omega\rho)

1675: \end{equation}

1676: The corresponding energy spectrum $E(\omega)=\omega n(\omega)$

1677: is shown in Fig.~\ref{fig_spectrum}.

1678: 

1679: \begin{figure}[h]

1680: \hspace*{-5mm}

1681: \begin{center}

1682: \epsfig{file=fig16.eps, width=70mm}

1683: \end{center}

1684: \caption{

1685: The energy spectrum $E(\omega)$ obtained from the analytic solution, 

1686: in units of $\frac{16}{g^2}(1-\kappa^2)^2$, versus $s=\omega\rho$.

1687: }\label{fig_spectrum}

1688: \end{figure}

1689: 

1690: Because the fourier transforms of the fields are finite, the occupation

1691: number (\ref{occno}) behaves as $1/\omega$ for small $\omega$.

1692: The number of particles is thus logarithmically divergent and should be

1693: cutoff at some low scale where the pQCD description of gluons is no longer 

1694: valid, leading to

1695: $

1696: N_g \sim  log( \frac{1}{ M_g \rho}) \,.

1697: $

1698: As explained in the previous Subsection, in heavy ion collisions

1699: this cutoff, identified with the gluon effective mass

1700: of Eq.~(\ref{Meff}), is still relatively small as compared to the

1701: typical momenta of gluons produced. 

1702: 

1703: The prompt particle energy distribution was also obtained in the numerical

1704: treatment of Section V, 

1705: defined as the integrand of the expression in Eq.~(\ref{penergy}),

1706: is shown in Fig.~\ref{edenfig} at a very late time in the evolution

1707: ($\tau = 50$ or $t\simeq 20$ fm).

1708: Like the previous distribution, it is finite at the origin, but in

1709: contrast it peaks at nonzero momentum.

1710: For illustrative purposes it is compared with a thermal distribution 

1711: of bosons at a temperature $T = 285$ MeV.

1712: We note that in an equilibrated environment,

1713: the effective mass and screening effects will only modify this profile below

1714: $k = 0.4$ GeV, where a relatively small part of the spectrum resides.

1715: 

1716: The produced gluons are free streaming, with no mechanism for equilibration,

1717: and yet our distribution is very similar to the thermal one for 

1718: momenta below about 1.5 GeV.

1719: Such a nearly thermal distribution has also been obtained from similar 

1720: calculations using an entirely different classical field approach 

1721: in Ref.~\cite{KV}.

1722: One can speculate that such an ostensible equilibration may contribute to the 

1723: the success of hydrodynamics in calculating particle spectra and elliptic 

1724: flow at RHIC \cite{TLS}.

1725: Our finding of four physical gluons from every decayed turning state is 

1726: also in line with RHIC entropy production, assuming that the density of

1727: these classical objects corresponds to that of instantons in the vacuum.

1728: The total energy of the turning state, found above to be 2.8 GeV, 

1729: is carried by gluons with a distribution peaked around 800 MeV.

1730: This average can be viewed as an upper bound, since in

1731: a more complete treatment a portion of this energy will be used for

1732: the production of light quark pairs.

1733: This next step will be addressed in a later publication.

1734: 

1735: \begin{figure}[bt]

1736: \begin{center}

1737: \epsfig{file=fig17.eps, width=80mm}

1738: \end{center}

1739: \caption{

1740: Energy spectrum of prompt gluons (solid line), obtained from the numerical 

1741: solution, and a thermal distribution with $T=285$ MeV (dashed line).

1742: }\label{edenfig}

1743: \end{figure}

1744: 

1745: \section{Conclusions}

1746: 

1747: In this work we have studied forced tunneling in pure SU(2) Yang-Mills theory.

1748: This process, generated through the excitation of instantons in the

1749: QCD vacuum, leads to unstable classical {\em turning states} which 

1750: explosively decay into gluonic radiation.

1751: These states and their decays are similar to the physics of sphalerons in

1752: the electroweak theory.

1753: 

1754: If this semi-classical treatment of pure Yang-Mills theory is indeed 

1755: relevant to the physics of QCD, the turning states should play a prominent

1756: role in the production of glue in high-energy hadronic collisions.

1757: In the case of $NN$ collisions the produced gluons will propagate into

1758: the QCD vacuum, to be quickly recombined into secondary hadrons.

1759: For heavy ion collisions, however, the large quantity of prompt glue 

1760: produced from the many turning states would be released into a highly 

1761: excited, perhaps deconfined medium.

1762: Observable consequences in both situations were discussed recently in

1763: Ref.~\cite{Shu_01}, and the results of this work support many of the

1764: estimates therein.

1765: Finally, we have obtained energy spectra for the prompt gluons that serve

1766: as the initial state in the dynamics of a heavy ion collision and found

1767: that, unlike the overall explosive dynamics, it is sensitive to the

1768: details of the initial turning state profile.

1769: We plan to further investigate phenomenological implications of the

1770: turning states in later works, with the role of fermion production a

1771: top priority.

1772: 

1773: {\bf Acknowledgements:}

1774: We thank I. Zahed and R. Venugopalan for useful discussions.

1775: This work is partially supported by the US-DOE grants Nos. DE-FG02-88ER40388

1776: and DE-FG03-97ER41014.

1777: 

1778: \begin{thebibliography}{99}

1779: \bibitem{BPST}

1780: A.~A.~Belavin, A.~M.~Polyakov, A.~S.~Shvarts, and Y.~S.~Tyupkin,

1781: %``Pseudoparticle Solutions Of The Yang-Mills Equations,''

1782: Phys.\ Lett.\ B {\bf 59}, 85 (1975).

1783: %%CITATION = PHLTA,B59,85;%%

1784: 

1785: \bibitem{tHooft}

1786: G.~'t~Hooft,

1787: %``Computation Of The Quantum Effects Due To A Four-Dimensional Pseudoparticle,''

1788: Phys.\ Rev.\ D {\bf 14}, 3432 (1976)

1789: [Erratum-ibid.\ D {\bf 18}, 2199 (1976)].

1790: %%CITATION = PHRVA,D14,3432;%%

1791: 

1792: \bibitem{inst_chiral}

1793: E.~V.~Shuryak,

1794: %``The Role Of Instantons In Quantum Chromodynamics. 1. Physical Vacuum,''

1795: Nucl.\ Phys.\ B {\bf 203}, 93 (1982);

1796: %%CITATION = NUPHA,B203,93;%%

1797: D.~Diakonov and V.~Y.~Petrov,

1798: %``A Theory Of Light Quarks In The Instanton Vacuum,''

1799: Nucl.\ Phys.\ B {\bf 272}, 457 (1986).

1800: %%CITATION = NUPHA,B272,457;%%

1801: 

1802: \bibitem{SS_98}

1803: T.~Schafer and E.~V.~Shuryak,

1804: %``Instantons in QCD,''

1805: Rev.\ Mod.\ Phys.\  {\bf 70}, 323 (1998).

1806: % [arXiv:hep-ph/9610451].

1807: %%CITATION = HEP-PH 9610451;%%

1808: 

1809: \bibitem{electroweak80s}

1810: V.~A.~Kuzmin, V.~A.~Rubakov, and M.~E.~Shaposhnikov,

1811: %``On The Anomalous Electroweak Baryon Number Nonconservation In The Early Universe,''

1812: Phys.\ Lett.\ B {\bf 155}, 36 (1985);

1813: %%CITATION = PHLTA,B155,36;%%

1814: P.~Arnold and L.~D.~McLerran,

1815: %``Sphalerons, Small Fluctuations And Baryon Number Violation In Electroweak Theory,''

1816: Phys.\ Rev.\ D {\bf 36}, 581 (1987).

1817: %%CITATION = PHRVA,D36,581;%%

1818: 

1819: \bibitem{electroweak90s}

1820: A.~Ringwald,

1821: Nucl.\ Phys.\ B {\bf 330}, 1 (1990);

1822: %%CITATION = NUPHA,B330,1;%%

1823: O.~Espinosa,

1824: %``High-Energy Behavior Of Baryon And Lepton Number Violating Scattering Amplitudes And Breakdown Of Unitarity In The Standard Model,''

1825: Nucl.\ Phys.\ B {\bf 343}, 310 (1990);

1826: %%CITATION = NUPHA,B343,310;%%

1827: V.~I.~Zakharov,

1828: %``Unitarity Constraints On Multiparticle Weak Production,''

1829: Nucl.\ Phys.\ B {\bf 353}, 683 (1991);

1830: %%CITATION = NUPHA,B353,683;%%

1831: M.~Maggiore and M.~A.~Shifman,

1832: Phys.\ Rev.\ D {\bf 46}, 3550 (1992);

1833: %%CITATION = PHRVA,D46,3550;%%

1834: D.~Diakonov and V.~Petrov,

1835: %``Nonperturbative isotropic multiparticle production in Yang-Mills theory,''

1836: Phys.\ Rev.\ D {\bf 50}, 266 (1994).

1837: %[arXiv:hep-ph/9307356].

1838: %%CITATION = HEP-PH 9307356;%%

1839: 

1840: \bibitem{KhR}

1841: V.~V.~Khoze and A.~Ringwald,

1842: Phys.\ Lett.\ B {\bf 259}, 106 (1991);

1843: %%CITATION = PHLTA,B259,106;%%

1844: 

1845: \bibitem{MRS}

1846: S.~Moch, A.~Ringwald, and F.~Schrempp,

1847: %``Instantons in deep-inelastic scattering: The simplest process,''

1848: Nucl.\ Phys.\ B {\bf 507}, 134 (1997).

1849: % [arXiv:hep-ph/9609445].

1850: %%CITATION = HEP-PH 9609445;%%

1851: 

1852: \bibitem{RS}

1853: A.~Ringwald and F.~Schrempp,

1854: %``Zooming-in on instantons at HERA,''

1855: Phys.\ Lett.\ B {\bf 503}, 331 (2001);

1856: % [arXiv:hep-ph/0012241].

1857: %%CITATION = HEP-PH 0012241;%%

1858: F.~Schrempp,

1859: %``Tracking QCD-instantons,''

1860: arXiv:hep-ph/0109032.

1861: %%CITATION = HEP-PH 0109032;%%

1862: 

1863: \bibitem{KKL} 

1864: D.~E.~Kharzeev, Y.~V.~Kovchegov, and E.~Levin,

1865: %``QCD instantons and the soft pomeron,''

1866: Nucl.\ Phys.\ A {\bf 690}, 621 (2001).

1867: %[arXiv:hep-ph/0007182].

1868: %%CITATION = HEP-PH 0007182;%%

1869: 

1870: \bibitem{NSZ}

1871: M.~A.~Nowak, E.~V.~Shuryak, and I.~Zahed,

1872: %``Instanton-induced inelastic collisions in QCD,''

1873: Phys.\ Rev.\ D {\bf 64}, 034008 (2001).

1874: % [arXiv:hep-ph/0012232].

1875: %%CITATION = HEP-PH 0012232;%%

1876: 

1877: \bibitem{COS} 

1878: G.~W.~Carter, D.~M.~Ostrovsky, and E.~V.~Shuryak,

1879: Phys.\ Rev.\ D {\bf 65}, 074034 (2002).

1880: % arXiv:hep-ph/0112036.

1881: %%CITATION = HEP-PH 0112036;%%

1882: 

1883: \bibitem{BFKL}

1884: E.~A.~Kuraev, L.~N.~Lipatov, and V.~S.~Fadin,

1885: %``The Pomeranchuk Singularity In Nonabelian Gauge Theories,''

1886: Sov.\ Phys.\ JETP {\bf 45} (1977) 199

1887: [Zh.\ Eksp.\ Teor.\ Fiz.\  {\bf 72} (1977) 377];

1888: %%CITATION = SPHJA,45,199;%%

1889: I.~I.~Balitsky and L.~N.~Lipatov,

1890: %``The Pomeranchuk Singularity In Quantum Chromodynamics,''

1891: Sov.\ J.\ Nucl.\ Phys.\  {\bf 28}, 822 (1978)

1892: [Yad.\ Fiz.\  {\bf 28}, 1597 (1978)];

1893: %%CITATION = SJNCA,28,822;%%

1894: L.~N.~Lipatov,

1895: %``The Bare Pomeron In Quantum Chromodynamics,''

1896: Sov.\ Phys.\ JETP {\bf 63}, 904 (1986)

1897: [Zh.\ Eksp.\ Teor.\ Fiz.\  {\bf 90}, 1536 (1986)].

1898: %%CITATION = SPHJA,63,904;%%

1899: 

1900: \bibitem{cgc}

1901: L.~D.~McLerran and R.~Venugopalan,

1902: %``Computing quark and gluon distribution functions for very large nuclei,''

1903: Phys.\ Rev.\ D {\bf 49}, 2233 (1994);

1904: % [arXiv:hep-ph/9309289].

1905: %%CITATION = HEP-PH 9309289;%%

1906: L.~D.~McLerran and R.~Venugopalan,

1907: %``Gluon distribution functions for very large nuclei at small transverse momentum,''

1908: Phys.\ Rev.\ D {\bf 49}, 3352 (1994);

1909: % [arXiv:hep-ph/9311205].

1910: %%CITATION = HEP-PH 9311205;%%

1911: L.~D.~McLerran and R.~Venugopalan,

1912: %``Green's functions in the color field of a large nucleus,''

1913: Phys.\ Rev.\ D {\bf 50}, 2225 (1994);

1914: % [arXiv:hep-ph/9402335].

1915: %%CITATION = HEP-PH 9402335;%%

1916: A.~Krasnitz and R.~Venugopalan,

1917: %``The initial energy density of gluons produced in very high energy  nuclear collisions,''

1918: Phys.\ Rev.\ Lett.\  {\bf 84}, 4309 (2000).

1919: % [arXiv:hep-ph/9909203].

1920: %%CITATION = HEP-PH 9909203;%%

1921: 

1922: \bibitem{KL}

1923: D.~Kharzeev and E.~Levin,

1924: %``Scale anomaly and 'soft' pomeron in QCD,''

1925: Nucl.\ Phys.\ B {\bf 578}, 351 (2000).

1926: %[arXiv:hep-ph/9912216].

1927: %%CITATION = HEP-PH 9912216;%%

1928: 

1929: \bibitem{Shu_toward}

1930: E.~V.~Shuryak,

1931: %``Toward the non-perturbative description of high energy processes,''

1932: Phys.\ Lett.\ B {\bf 486}, 378 (2000).

1933: % [arXiv:hep-ph/0001189].

1934: %%CITATION = HEP-PH 0001189;%%

1935: 

1936: \bibitem{Manton}

1937: N.~S.~Manton,

1938: %``Topology In The Weinberg-Salam Theory,''

1939: Phys.\ Rev.\ D {\bf 28}, 2019 (1983);

1940: %%CITATION = PHRVA,D28,2019;%%

1941: 

1942: \bibitem{KM}

1943: F.~R.~Klinkhamer and N.~S.~Manton,

1944: %``A Saddle Point Solution In The Weinberg-Salam Theory,''

1945: Phys.\ Rev.\ D {\bf 30}, 2212 (1984).

1946: %%CITATION = PHRVA,D30,2212;%%

1947: 

1948: \bibitem{Witten_ans}

1949: E.~Witten,

1950: %``Some Exact Multipseudoparticle Solutions Of Classical Yang-Mills  Theory,''

1951: Phys.\ Rev.\ Lett.\  {\bf 38}, 121 (1977).

1952: %%CITATION = PRLTA,38,121;%%

1953: 

1954: \bibitem{AKY} 

1955: T.~Akiba, H.~Kikuchi, and T.~Yanagida,

1956: %``Static Minimum Energy Path From A Vacuum To A Sphaleron In The Weinberg-Salam Model,''

1957: Phys.\ Rev.\ D {\bf 38}, 1937 (1988).

1958: %%CITATION = PHRVA,D38,1937;%%

1959: 

1960: \bibitem{FKS}  

1961: E.~Farhi, V.~V.~Khoze, and R.~J.~Singleton,

1962: %``Minkowski space nonAbelian classical solutions with noninteger winding number change,''

1963: Phys.\ Rev.\ D {\bf 47}, 5551 (1993).

1964: % [arXiv:hep-ph/9212239].

1965: %%CITATION = HEP-PH 9212239;%%

1966: 

1967: \bibitem{CDG}

1968: C.~G.~Callan, R.~F.~Dashen, and D.~J.~Gross,

1969: %``Toward A Theory Of The Strong Interactions,''

1970: Phys.\ Rev.\ D {\bf 17}, 2717 (1978).

1971: %%CITATION = PHRVA,D17,2717;%%

1972: 

1973: % \bibitem{DP_ridge}

1974: % D.~Diakonov, M.~V.~Polyakov, P.~Sieber, J.~Schaldach, and K.~Goeke,

1975: % %``Fermion sea along the sphaleron barrier,''

1976: % Phys.\ Rev.\ D {\bf 49}, 6864 (1994).

1977: % % [arXiv:hep-ph/9311374].

1978: % %%CITATION = HEP-PH 9311374;%%

1979: 

1980: \bibitem{yung}A.~V.~Yung,

1981: %``Instanton Vacuum In Supersymmetric QCD,''

1982: Nucl.\ Phys.\ B {\bf 297}, 47 (1988).

1983: %%CITATION = NUPHA,B297,47;%%

1984: 

1985: \bibitem{SS96}

1986: T.~Schafer and E.~V.~Shuryak,

1987: %``The instanton liquid in QCD at zero and finite temperature,''

1988: Phys.\ Rev.\ D {\bf 53}, 6522 (1996).

1989: % [arXiv:hep-ph/9509337].

1990: %%CITATION = HEP-PH 9509337;%

1991: 

1992: \bibitem{Shu_rat}

1993: E.~V.~Shuryak,

1994: Nucl.\ Phys.\ B {\bf 302}, 574 (1988).

1995: %%CITATION = NUPHA,B302,574;%%

1996: 

1997: \bibitem{Shu_qm}

1998: E.~V.~Shuryak,

1999: Nucl.\ Phys.\ B {\bf 302}, 621 (1988).

2000: %%CITATION = NUPHA,B302,621;%%

2001: 

2002: \bibitem{Verbaarschot}J.~J.~Verbaarschot,

2003: %``Streamlines and conformal invariance in Yang-Mills theories,''

2004: Nucl.\ Phys.\ B {\bf 362}, 33 (1991)

2005: [Erratum-ibid.\ B {\bf 386}, 236 (1991)].

2006: %%CITATION = NUPHA,B362,33;%%

2007: 

2008: \bibitem{Smilga}

2009: A.~V.~Smilga,

2010: %``Sphalerons, instantons, and standing waves on S**3 x R,''

2011: Nucl.\ Phys.\ B {\bf 459}, 263 (1996).

2012: % [arXiv:hep-th/9504117].

2013: %%CITATION = HEP-TH 9504117;%%

2014: 

2015: \bibitem{Luscher}

2016: M.~L\"uscher,

2017: %``SO(4) Symmetric Solutions Of Minkowskian Yang-Mills Field Equations,''

2018: Phys.\ Lett.\ B {\bf 70}, 321 (1977).

2019: %%CITATION = PHLTA,B70,321;%%

2020: 

2021: \bibitem{Schechter}

2022: B.~M.~Schechter,

2023: %``Yang-Mills Theory On The Hypertorus,''

2024: Phys.\ Rev.\ D {\bf 16}, 3015 (1977).

2025: 

2026: \bibitem{HK}

2027: M.~Hellmund and J.~Kripfganz,

2028: %``The Decay of the sphalerons,''

2029: Nucl.\ Phys.\ B {\bf 373}, 749 (1992).

2030: %%CITATION = NUPHA,B373,749;%%

2031: 

2032: \bibitem{Zad}

2033: J.~Zadrozny,

2034: %``Sphaleron decay products: A Coherent state analysis,''

2035: Phys.\ Lett.\ B {\bf 284}, 88 (1992).

2036: %%CITATION = PHLTA,B284,88;%%

2037: 

2038: \bibitem{AKY2} 

2039: T.~Akiba, H.~Kikuchi, and T.~Yanagida,

2040: %``The Free Energy Of The Sphaleron In The Weinberg-Salam Model,''

2041: Phys.\ Rev.\ D {\bf 40}, 588 (1989).

2042: %%CITATION = PHRVA,D40,588;%%

2043: 

2044: \bibitem{Shu_80}

2045: E.~V.~Shuryak,

2046: %``Quantum Chromodynamics And The Theory Of Superdense Matter,''

2047: Phys.\ Rept.\  {\bf 61}, 71 (1980).

2048: %%CITATION = PRPLC,61,71;%%

2049: 

2050: \bibitem{LH}

2051: P.~Levai and U.~W.~Heinz,

2052: %``Massive gluons and quarks and the equation of state obtained from SU(3)  lattice QCD,''

2053: Phys.\ Rev.\ C {\bf 57}, 1879 (1998).

2054: %[arXiv:hep-ph/9710463].

2055: %%CITATION = HEP-PH 9710463;%%

2056: 

2057: \bibitem{TW}

2058: J.~C.~Taylor and S.~M.~Wong,

2059: %``The Effective Action Of Hard Thermal Loops In QCD,''

2060: Nucl.\ Phys.\ B {\bf 346}, 115 (1990).

2061: %%CITATION = NUPHA,B346,115;%%

2062: 

2063: \bibitem{KV}

2064: A.~Krasnitz and R.~Venugopalan,

2065: %``The initial gluon multiplicity in heavy ion collisions,''

2066: Phys.\ Rev.\ Lett.\  {\bf 86}, 1717 (2001).

2067: % [arXiv:hep-ph/0007108].

2068: %%CITATION = HEP-PH 0007108;%%

2069: 

2070: \bibitem{TLS}

2071: D.~Teaney, J.~Lauret, and E.~V.~Shuryak,

2072: %``A hydrodynamic description of heavy ion collisions at the SPS and RHIC,''

2073: arXiv:nucl-th/0110037.

2074: %%CITATION = NUCL-TH 0110037;%%

2075: 

2076: \bibitem{Shu_01}E.~V.~Shuryak,

2077: Phys.\ Lett.\ B {\bf 515}, 359 (2001).

2078: % [hep-ph/0101269].

2079: %%CITATION = HEP-PH 0101269;%

2080: 

2081: % \bibitem{KKV}

2082: % D.~Kharzeev, A.~Krasnitz, and R.~Venugopalan,

2083: % %``Anomalous chirality fluctuations in the initial stage of heavy ion  collisions and parity odd bubbles,''

2084: % arXiv:hep-ph/0109253.

2085: % %%CITATION = HEP-PH 0109253;%%

2086: 

2087: \end{thebibliography}

2088: \end{document}

2089: