1: % PACS number: 0460P

2: % Keywords: Quantum Geometry, Loop Quantum Cosmology

3: %\documentstyle[psfrag,preprint,graphicx,tighten,eqsecnum,floats,aps,amssymb]{revtex}

4: %\documentstyle[psfrag,preprint,graphicx,tighten,floats,aps]{revtex}

5: %\documentclass[preprint,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}

6: \documentclass[preprint,showpacs,preprintnumbers,amsmath,amssymb,nofootinbib]{revtex4}

7: \usepackage{amssymb}

8: \usepackage[mathscr]{eucal}

9: \usepackage{graphicx}

10: \def\be{\begin{equation}}

11: \def\ee{\end{equation}}

12: \def\ba{\begin{eqnarray}}

13: \def\ea{\end{eqnarray}}

14: 

15: \def\C{{\cal C}}

16: \def\M{{\cal M}}

17: \def\Gauss{{\cal C}_{\rm Gauss}}

18: \def\ham{{\cal C}_{\rm Ham}}

19: \def\H{{\cal H}}

20: \def\Ab{{\bar{{\cal A}}_{\rm KS}}}

21: \def\cyl{{\rm Cyl_{\rm KS}}}

22: \def\cylstar{{\rm Cyl}^\star_{\rm KS}}

23: \def\tr{{\rm Tr\,}}

24: \def\Tr{{\rm Tr\,}}

25: \def\su{{\rm su}}

26: \def\SU{{\rm SU}}

27: 

28: \def\lp{{\ell}_{\rm Pl}}

29: \def\q{{}^o\!q}

30: \def\e{{}^o\!e}

31: \def\w{{}^o\!\omega}

32: \def\L{\delta}

33: \def\f{\frac}

34: \def\T{T}

35: \def\nu{\tau}

36: 

37: \newcommand{\md}{{\mathrm d}}

38: \newcommand{\vt}{\vartheta}

39: \newcommand{\vp}{\varphi}

40: \newcommand{\R}{{\mathbb R}}

41: \newcommand{\sgn}{{\mathrm{sgn}}}

42: \def\S{{\mathbb S}}

43: \def\F{{}^o\!F}

44: 

45: %\newcommand{\tr}{{\mathrm{tr}}}

46: %\newcommand{\diag}{{\mathrm{diag}}}

47: %\newcommand{\case}[2]{{\textstyle \frac{#1}{#2}}}

48: %\newcommand{\lP}{\ell_{\rm P}}

49: %\newcommand{\rip}[2]{\ensuremath{(#1\,|\,#2\rangle}}

50: %\newcommand{\norm}[1]{\ensuremath{\|#1\|}}

51: %\newcommand{\bra}[1]{\ensuremath{\langle#1|}}

52: %\newcommand{\ket}[1]{\ensuremath{|#1\rangle}}

53: %\newcommand{\ip}[2]{{\langle#1\,|\,#2\rangle}}

54: 

55: 

56: 

57: \begin{document}

58: %\date\today

59: 

60: \preprint{IGPG--05--09/01,AEI--2005--132}

61: \title{Quantum geometry and the Schwarzschild singularity}

62: \author{Abhay\ Ashtekar${}^{1,2}$ and Martin Bojowald${}^{2,1}$}

63: \address{1. Institute for Gravitational Physics and Geometry,\\

64: Physics Department, Penn State, University Park, PA 16802, USA\\

65: 2. Max-Planck-Institut f\"ur Gravitationsphysik, Albert-Einstein-Institut,

66: Am M\"uhlenberg 1, D-14476 Potsdam, Germany}

67: 

68: 

69: \begin{abstract}

70: 

71: In homogeneous cosmologies, quantum geometry effects lead to a

72: resolution of the classical singularity without having to invoke

73: special boundary conditions at the singularity or introduce ad-hoc

74: elements such as unphysical matter. The same effects are shown to

75: lead to a resolution of the Schwarzschild singularity. The

76: resulting quantum extension of space-time is likely to have

77: significant implications to the black hole evaporation process.

78: Similarities and differences with the situation in quantum

79: geometrodynamics are pointed out.

80: 

81: \end{abstract}

82: 

83: 

84: \pacs{04.60.Pp, 04.70.Dy, 04.60.Nc, 03.65.Sq}

85: \maketitle

86: 

87: 

88: \section{Introduction}

89: \label{s1}

90: 

91: General relativity provides a subtle and powerful interplay

92: between gravity and geometry, thereby opening numerous

93: possibilities for novel phenomena. Among the most spectacular of

94: the resulting conceptual advances are the predictions that the

95: universe began with a big bang and massive stars can end their

96: lives as black holes. In both cases, one encounters singularities.

97: Space-time of general relativity literally ends and classical

98: physics comes to a halt.

99: 

100: However, general relativity is incomplete because it ignores

101: quantum effects. It is widely believed that quantum gravity

102: effects become significant in the high curvature regions that

103: develop {before} singularities are formed. These are likely to

104: significantly change the space-time structure, making the

105: predictions of general relativity unreliable. Hence, real physics

106: need not stop at the big bang and black hole singularities. While

107: the classical space-time does end there, quantum space-time may

108: well extend beyond.

109: 

110: Attempts at extending physics beyond the big-bang singularity date

111: back at least to the seventies when mini-superspaces were

112: introduced and quantum cosmology was born (see, e.g., \cite{cm}).

113: These investigations sparked off new developments in different

114: directions ranging from foundations of quantum mechanics to the

115: development of WKB methods to test semi-classicality in quantum

116: cosmology, to the introduction of novel Euclidean methods to

117: calculate the appropriate wave function of the universe. This

118: deepened our understanding of issues related to the quantum

119: physics of the universe as a whole \cite{jh-lh}. However, in these

120: approaches the singularity was not generically resolved. A quantum

121: extension of space-time either required the introduction of new

122: principles \cite{hh}, or ad-hoc assumptions, such as existence of

123: matter violating energy conditions already at the classical level,

124: or of external clocks that remain insensitive to the infinite

125: curvature encountered at singularities.

126: 

127: Emergence of quantum Riemannian geometry and the associated

128: mathematical techniques \cite{alrev,crbook,ttrev} have provided a

129: new approach to revisit the issue of quantum extensions. Within

130: the mini-superspaces used in quantum cosmology, the big-bang

131: singularity could be resolved \emph{without} having to introduce

132: external boundary conditions at the big-bang, or matter/clocks

133: with unphysical properties \cite{mb3,mb6,abl}. In the homogeneous,

134: isotropic mini-superspace coupled to a massless scalar field, in

135: particular, an exhaustive analysis can be carried out \cite{aps}.

136: At the analytic level, one can introduce an appropriate inner

137: product on physical states, define Dirac observables and, using

138: the two, construct semi-classical states. One can then use

139: numerical methods to examine the nature of the quantum space-time.

140: If one begins with a semi-classical state at large late times

141: (`now') and evolves it back in time, it remains semi-classical

142: till one encounters the deep Planck regime near the classical

143: singularity. In this regime the quantum geometry effects dominate.

144: However, the state becomes semi-classical again on the other side;

145: the deep Planck region serves as a quantum bridge between two

146: large, classical space-times. The model is too simple to be

147: applied reliably to the actual universe we live in. However at a

148: conceptual and mathematical level it demonstrates the new

149: possibilities, systematically made available by quantum geometry.

150: 

151: It is then natural to ask if a similar resolution of singularities

152: occurs also in the context of black holes. Now, there is an

153: extensive literature on the nature of black hole singularities in

154: the classical theory. In particular, detailed mathematical

155: analysis of spherical collapse of uncharged matter was carried out

156: by Christodoulou, Dafermos and others. It has led to the

157: expectation that, for black holes formed through gravitational

158: collapse, the singularity would be generically space-like (see,

159: e.g., \cite{md} and references therein). Therefore, it is natural

160: to focus on this case first and ask for the nature of quantum

161: space-time that would result via their resolution.%

162: %

163: \footnote{Indeed, one does not expect quantum gravity effects to

164: resolve all singularities. If, for example, they led to a

165: resolution of the time-like singularity of the negative energy

166: Schwarzschild space-time, energy would be unbounded below in

167: quantum gravity and the theory would have unphysical features

168: \cite{hm}. Rather, one expects that there would be \emph{no}

169: physical states in quantum gravity that would resemble negative

170: energy Schwarschild geometry at large distances, whence the issue

171: of `resolution' would simply not arise.}

172: %

173: If the singularity is resolved and {if} the quantum geometry in

174: the deep Planck regime again serves as a bridge to a large

175: classical region beyond, there would be no information loss in the

176: black hole formation and evaporation process \cite{ab2} (see also

177: \cite{s'thw,sh3}). Pure states in the distant past would evolve to

178: pure states in the distant future and one would have a space-time

179: description of the entire process in a quantum mechanical setting.

180: Thus, a detailed analysis of the fate of black hole singularities

181: is of considerable importance also for the fundamental issue of

182: whether the standard unitary evolution of quantum physics has to

183: be modified in the setting of black holes.

184: 

185: The purpose of this note is to initiate this investigation using

186: quantum geometry, in the setting of connection dynamics. We will

187: focus just on the mini-superspace that is appropriate for

188: describing the geometry interior to the horizon of a Schwarzschild

189: black hole. Since this mini-superspace is spatially homogeneous

190: ---of Kantowski-Sachs type--- one can take over the techniques

191: that have been developed in the setting of homogeneous cosmologies

192: \cite{mb6}. Our main result is that the quantum scalar constraint

193: is indeed such that the singularity is resolved. The salient

194: features of our analysis are as follows. First, we use a

195: \emph{self-adjoint} Hamiltonian constraint. Therefore, our

196: analysis can serve as the point of departure for the construction

197: of the physical Hilbert space either via the group averaging

198: procedure \cite{group} \emph{or} via deparametrization of the

199: theory as in \cite{aps}. Second, we spell out the symmetry

200: reduction procedure that leads to the Kantowski-Sachs

201: mini-superspace \emph{within connection dynamics}. We will see in

202: sections \ref{s3} and \ref{s4} that the structure of the resulting

203: phase space is important to the issue of singularity resolution.%

204: %

205: \footnote{A quantization of the Kantowski-Sachs minisuperspace is

206: available in the geometrodynamical framework, \cite{m}.

207: Unfortunately, it does not shed light on singularity resolution

208: because the location of the singularity in the minisuperspace was

209: misidentified. In the notation used in \cite{m}, the singularity

210: lies at $(a=\infty, b=0)$ ---not at $(a=0,b=0)$ as assumed there.

211: Therefore, neither is the inverse volume operator (17) of \cite{m}

212: bounded or well-defined at the singularity nor does the discrete

213: equation (23) of \cite{m} enable one to evolve across the

214: singularity.}

215: %

216: In particular, this space is an extension of the phase space used

217: in geometrodynamics in that one allows for degenerate triads (and

218: hence 3-metrics). Thanks to this enlargement, points of the

219: reduced phase space corresponding to the singularity do not

220: constitute a boundary, whence the support of the wave function

221: `beyond singularity' can be interpreted geometrically. However, we

222: emphasize that the results of this note constitute only initial

223: steps for a more complete theory, e.g., along the lines of

224: \cite{aps}.

225: 

226: Our discussion is organized as follows. In section \ref{s2} we

227: discuss the classical theory of the Kantowski-Sachs

228: mini-superspace using connection-dynamics. In section \ref{s3} we

229: analyze the kinematics of its quantum theory and in section

230: \ref{s4} we present quantum dynamics. Section \ref{s5} provides a

231: brief summary and directions for future work.

232: 

233: \section{Classical Theory}

234: \label{s2}

235: 

236: This section is divided into two parts. In the first, we carry out

237: the Kantowski-Sachs symmetry reduction of vacuum general

238: relativity and in the second we discuss the structure of the

239: resulting phase space.

240: 

241: \subsection{Symmetry reduction}

242: \label{s2.1}

243: 

244: The portion of the space-time interior to the horizon of a

245: Schwarzschild black hole can be naturally foliated by  3-manifolds

246: $M$ with topology $\R\times \S^2$ such that the geometry on each

247: slice is invariant under the Kantowski--Sachs symmetry group

248: $G=\R\times {\rm SO(3)}$. In the discussion of this model we

249: closely follow \cite{abl} where the mathematical structure of the

250: simpler, isotropic model is studied in detail.

251: 

252: As in any homogeneous situation, $M$ acquires an equivalence class

253: of positive definite metrics related by an overall constant. We

254: fix one such metric $\q_{ab}$ as well as an orthonormal triad

255: $\e_i^a$ and co-triad $\w_a^i$, compatible with it. With the given

256: symmetry, we choose the metric on the orbits of the ${\rm SO(3)}$

257: action to be the unit 2-sphere metric,

258: $\md\vt^2+\sin^2\vt\md\vp^2$ in polar coordinates $\vt,\vp$.

259: Locally, the corresponding co-triad elements $\md\vt$, and

260: $\sin\vt\md\vp$ can be completed to an orthonormal co-triad by a

261: third form $\alpha$.  Thanks to the Maurer--Cartan relations for

262: the symmetry group, $\alpha$ must be closed and therefore exact on

263: the given topology, whence it defines a third adapted coordinate

264: $x$ via $\alpha=\md x$. In these coordinates the background metric

265: has line element

266: %

267:  $$\md s_o^2= \md x^2+\md\vt^2+\sin^2\vt\md\vp^2$$

268: %

269: and determinant $\q=\sin^2\vt$.

270: 

271: In connection dynamics, the canonically conjugate pair consists of

272: fields $(A_a^i, E^a_i)$ of an ${\rm SU(2)}$ connection 1-form

273: $A_a^i$ and a (possibly degenerate) triad $E^a_i$ of density

274: weight 1 on $M$. This pair is invariant under the symmetry group

275: if it satisfies:

276: %

277: \be \mathcal{L}_\xi\, A = D\Lambda \quad {\rm and} \quad

278: \mathcal{L}_\xi E = [E, \Lambda] \label{sym} \ee

279: %

280: for any symmetry vector field $\xi^a$ and some generator

281: $\Lambda^i$ of local, ${\rm SU(2)}$ gauge transformations (which

282: may depend on $\xi^a$). One can verify that any such symmetric

283: pair is gauge equivalent to fields of the form:%

284: %

285:  \footnote{More precisely, invariant connections \cite{kn,b}

286: carry a non-negative, integer-valued topological charge which we

287: have set equal to one in the above expressions. (In general, it

288: would multiply the last contribution to (\ref{sym1}); see, e.g.,

289: \cite{bk}.) Each value of this charge gives rise to an independent

290: sector of invariant connections. However, only the sector used

291: here allows non-degenerate triads invariant under (\ref{sym}) with

292: {\em the same} local gauge transformation in $A$ and $E$. That $A$

293: in (\ref{sym1}) satisfies (\ref{sym}) can be verified as follows.

294: If we denote the three generators of the ${\rm SO(3)}$ action on

295: $M$ by

296:  $X=\sin\vp\partial_{\vt}+\cot\vt\cos\vp\partial_{\vp}$,

297:  $Y=-\cos\vp\partial_{\vt}+\cot\vt\sin\vp\partial_{\vp}$, and

298:  $Z=\partial_{\vp}$, and the generator of the $\R$-action by $t^a$,

299: then $\mathcal{L}_X A = D(\cos\vp/\sin\vt)\tau_3$,\,

300: $\mathcal{L}_Y A = D (\sin\vp/\sin\vt)\tau_3$,\, $\mathcal{L}_Z A

301: = 0$ and $\mathcal{L}_t A = 0$. Invariance of $E$ in (\ref{sym2})

302: follows analogously, or by noting that $E^a_i\delta A_a^i$ defines

303: a gauge invariant and (spatially) constant 1-form on phase space.}

304: %

305: \ba A &=& \tilde{c}\tau_3\md

306: x+(\tilde{a}\tau_1+\tilde{b}\tau_2)\md\vt

307: +(-\tilde{b}\tau_1+\tilde{a}\tau_2)\sin\vt\md\vp+

308: \tau_3 \cos\vt \md\vp \label{sym1}\\

309: E &=& \tilde{p}_c\tau_3\sin\vt\frac{\partial}{\partial

310: x}+(\tilde{p}_a\tau_1+\tilde{p}_b\tau_2)

311: \sin\vt\frac{\partial}{\partial\vt}

312: +(-\tilde{p}_b\tau_1+\tilde{p}_a\tau_2)\frac{\partial}{\partial\vp}\,.

313: \label{sym2} \ea

314: %

315: where $\tilde{a},\tilde{b},\tilde{c}$ are real constants and

316: $\tau^j$ is the standard basis in ${\rm su(2)}$, satisfying

317: $[\tau_i,\, \tau_j] = \epsilon_{ij}{}^k \tau_k$. However, this

318: does not exhaust the freedom to perform local ${\rm SU(2)}$ gauge

319: transformations entirely: There is still the freedom to perform a

320: \emph{global} $\SU(2)$ transformation along $\tau_3$ which rotates

321: the `vectors' $(\tilde{a}, \tilde{b})$ and $(\tilde{p_a},

322: \tilde{p_b})$. We will return to this freedom in the next

323: sub-section.

324: 

325: From the invariant triad density (\ref{sym2}) we can derive the

326: corresponding co-triad $\omega$:

327: %

328: \be\label{cotriad} \omega = \omega_c \tau_3 \md x +

329: (\omega_a\tau_1 + \omega_b\tau_2) \md \vt + (-\omega_b \tau_1 +

330: \omega_a \tau_2) \sin\vt \md\vp\ee

331: %

332: where

333: %

334: \be \omega_c = \frac{{\rm sgn}\tilde{p}_c\,

335: \sqrt{\tilde{p}^2_a+\tilde{p}^2_b}}{\sqrt{|\tilde{p}_c|}};\quad

336:  \omega_b = \frac{\sqrt{|\tilde{p}_c|}\, \tilde{p}_b}

337: {\sqrt{\tilde{p}^2_a+\tilde{p}^2_b}}; \quad {\rm and} \,\,

338: \omega_a = \frac{\sqrt{|\tilde{p}_c|}\,\tilde{p}_a}

339: {\sqrt{\tilde{p}^2_a+\tilde{p}^2_b}}\, . \ee

340: %

341: The co-triad in turn determines its spin-connection, the ${\rm

342: su(2)}$-valued 1-form $\Gamma$. As one might expect from the

343: structure of the triad, $\Gamma$ turns out to be the standard

344: `magnetic monopole' connection:

345: %

346: \begin{equation} \label{Gamma}

347:  \Gamma= \cos\vt\tau_3\, \md\vp\,.

348: \end{equation}

349: %

350: Consequently, the ${\rm su(2)}$-valued extrinsic curvature 1-form

351: $K$ is given by:

352: %

353: \begin{equation} \label{K}

354:  \gamma K : =A - \Gamma= \tilde{c}\tau_3\md x +

355: (\tilde{a}\tau_1+\tilde{b}\tau_2)\md\vt

356:  + (-\tilde{b}\tau_1+\tilde{a}\tau_2)\sin\vt\md\vp\,.

357: \end{equation}

358: %

359: Finally the curvature $F_{ab}$ of $A_a$ can be easily computed.

360: Its dual, the ${\rm su(2)}$-valued vector density $B^a\, :=\,

361: \frac{1}{2} \eta^{abc}F_{bc}$ has the same invariant form as $E^a$:

362: %

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

364:  B=(\tilde{a}^2+\tilde{b}^2-1)\tau_3\sin\vt\frac{\partial}{\partial

365:  x}+(\tilde{a}\tau_1+\tilde{b}\tau_2)\tilde{c}

366: \sin\vt\frac{\partial}{\partial\vt}

367:  +(-\tilde{b}\tau_1+\tilde{a}\tau_2)\tilde{c}\frac{\partial}{\partial\vp}

368: \end{equation}

369: %

370: These expressions will be used in sections \ref{s3} and \ref{s4}.

371: 

372: 

373: \subsection{The reduced phase space}

374: \label{s2.2}

375: 

376: The symmetry reduction procedure led us to a 6-dimensional reduced

377: phase space $\tilde{\bf \Gamma}$ with coordinates $(\tilde{a},

378: \tilde{b}, \tilde{c};\, \tilde{p}_a, \tilde{p}_b, \tilde{p}_c)$.

379: Let us begin by computing the symplectic structure $\tilde{\bf

380: \Omega}$ on it. The basic idea, as in all symmetry reductions, is

381: to pull-back the symplectic structure of the full theory to the

382: symmetry reduced phase space. However, since the full symplectic

383: structure involves an integral over $M$ and since the fields of

384: interest are homogeneous also in the non-compact

385: $\mathbb{R}$-direction, as usual, we are led to consider only a

386: finite interval $\cal{I}$ in the $\mathbb{R}$ direction. Let the

387: length of this interval (w.r.t. the fiducial metric $\q_{ab}$) be

388: $L_o$. Then, the symplectic structure on the reduced phase space

389: is given by:

390: %

391: \begin{equation}

392: \tilde{\bf \Omega} = \frac{L_o}{2\gamma G}(2\md\tilde{a}\wedge\md

393: \tilde{p}_a+ 2\md\tilde{b}\wedge\md \tilde{p}_b+ \md

394: \tilde{c}\wedge\md \tilde{p}_c)

395: \end{equation}

396: %

397: where $\gamma$ is the Barbero--Immirzi parameter and $G$ the

398: gravitational constant. (The volume of the $\S^2$ orbits of the

399: ${\rm SO(3)}$ action, defined by the fiducial metric, is the standard

400: one, $4\pi$.)

401: 

402: Because of the form (\ref{sym1}) and (\ref{sym2}) of invariant

403: connections and triads, the vector constraint is automatically

404: satisfied. However, because of the residual, global ${\rm SU(2)}$

405: gauge freedom mentioned in section \ref{s2.1}, the Gauss

406: constraint is not. Inserting the invariant connection and triad

407: into the Gauss constraint we obtain

408: %

409: \begin{equation}

410:  \Gauss=\tilde{a}\tilde{p}_b-\tilde{b}\tilde{p}_a=0

411: \end{equation}

412: %

413: which generates simultaneous rotations of the pairs

414: $(\tilde{a},\tilde{b})$ and $(\tilde{p}_a,\tilde{p}_b)$. Thus,

415: only the `norms' $\sqrt{\tilde{a}^2+\tilde{b}^2}$,

416: $\sqrt{\tilde{p}_a^2+\tilde{p}_b^2}$ and the `scalar product'

417: $(\tilde{a}\tilde{p}_a+\tilde{b}\tilde{p}_b)$ are gauge invariant.

418: We will fix this gauge freedom as follows. If

419: $(\tilde{a},\tilde{b})=0$, we rotate the triad components such

420: that $\tilde{p}_a=0$. Otherwise we rotate $(\tilde{a},\tilde{b})$

421: such that $\tilde{a}=0$ which implies $\tilde{b}\not=0$. Then,

422: $\Gauss=0$ implies $\tilde{p}_a=0$. There is still a residual

423: gauge freedom\,\, $\Pi_b\colon (b, p_b) \rightarrow (-b, -p_b)$\,\,

424: which changes the signs of $b$ and $p_b$ simultaneously. This is

425: just the parity transformation in the $p_b$ variable. One can

426: either retain this freedom and ensure that all the final

427: constructions are invariant with respect to this `$b$-reflection'

428: \emph{or} eliminate it by a gauge choice such as $\tilde{p}_b \ge

429: 0$. This gauge choice turns out to be particularly convenient for

430: classical dynamics. In the quantum theory, on the other hand, it

431: is more natural to retain the freedom at first and then ask that

432: physical states be invariant under the parity operator

433: $\hat{\Pi}_b$ implementing this transformation. Therefore, we will

434: allow both possibilities. In either case the 4-dimensional phase

435: space carries a global chart $(\tilde{b},\tilde{c};

436: \tilde{p}_b,\tilde{p}_c)$ (where $\tilde{p}_b \ge0$ if the

437: `$b$-parity gauge' is fixed.) These coordinates are subject to the

438: scalar or the Hamiltonian constraint, discussed below.

439: 

440: We first note that these variables and, because of its explicit

441: dependence on $L_o$, the symplectic structure $\tilde{\bf \Omega}$

442: depend on the fiducial metric. It is convenient to remove this

443: dependence by rescaling the variables in a manner that is

444: motivated by their scaling properties,

445: %

446: \begin{equation}

447:  (b,c):=(\tilde{b},L_o\tilde{c})\quad,\quad

448:  (p_b,p_c):=(L_o\tilde{p}_b,\tilde{p}_c)\,.

449: \end{equation}

450: %

451: We now have a one-to-one parametrization of the gauge fixed phase

452: space ${\bf \Gamma} $ by $(b,p_b,c,p_c)$ with symplectic structure

453: %

454: \begin{equation} \label{symp}

455: {\bf \Omega} = \frac{1}{2\gamma G}(2\md b\wedge\md p_b+\md

456: c\wedge\md p_c)\,.

457: \end{equation}

458: %

459: For later purposes let us express the volume of our elementary

460: cell ${\cal I}\times \S^2$ and areas bounded by our preferred

461: family of curves as functions on the reduced phase space. The

462: volume is given by:

463: %

464: \begin{equation} \label{vol}

465:  V=\int\md^3x\sqrt{\left|\det E\right|}=4\pi L_o

466:  \sqrt{|\tilde{p}_c|} |\tilde{p}_b| = 4\pi\sqrt{|p_c|} |p_b|\,.

467: \end{equation}

468: %

469: The three surfaces $S_{x,\vp}$, $S_{x,\vt}$ and $S_{\vt,\vp}$ of

470: interest are respectively bounded by the interval ${\cal I}$ and

471: the equator, ${\cal I}$ and a great circle along a longitude, and

472: the equator and a longitude (so that $S_{\vt,\vp}$ forms a quarter

473: of the sphere $\S^2$). Their areas are given by:

474: %

475: \be \label {area}

476:  A_{x, \vt} = 2\pi |p_b|, \quad A_{x, \vp} = 2 \pi |p_b|,

477:  \quad{\rm and} \quad  A_{\vt,\vp} = \pi |p_c|\,. \ee

478: 

479: Finally, let us consider the Hamiltonian constraint and classical

480: dynamics. On the reduced phase space ${\bf \Gamma}$, the

481: constraint functional of the full theory

482: %

483: \begin{equation} \label{fullham}

484: \ham= \int\md x^3 N e^{-1}\, [\epsilon_{ijk} F_{ab}^i E^a_jE^b_k-

485: 2(1+\gamma^2) K_{[a}^iK_{b]}^j E_i^aE_j^b]\, ,

486: \end{equation}

487: %

488: where $e:= \sqrt{\left|\det E\right|}\,{\rm sgn}(\det E)$, reduces

489: to

490: %

491: \begin{eqnarray} \label{ksham}

492: \ham = -\frac{8\pi N}{\gamma^2}\frac{\sgn p_c}{\sqrt{|p_c|}p_b}

493: \left[(b^2+\gamma^2)p_b^2+

494:   2cp_cbp_b\right]\, ,

495: \end{eqnarray}

496: %

497: where, as is usual in the homogeneous models, we have chosen a

498: constant lapse function $N$. To simplify the equations of motion,

499: it is convenient%

500: %

501: \footnote{With this choice, the points $b=0$ or $p_c =0$ have to

502: be excised in the discussion of dynamics. The explicit form of the

503: solutions below shows that this is not overly restrictive since on

504: classical solutions these points correspond to the singularity or

505: the horizon.}

506: %

507: to choose $N = \frac{\gamma {\rm sgn}p_c\, \sqrt{|p_c|}}{16\pi G

508: b}$. Then the Hamiltonian constraint becomes $\ham:=-(2\gamma

509: G)^{-1}[(b^2+\gamma^2)p_b/b+2cp_c]$, yielding the following

510: equations of motion:

511: %

512: \begin{eqnarray*}

513:  \dot{b} = -\frac{1}{2}(b+\gamma^2b^{-1}), &\quad\quad&

514:  \dot{p}_b = \frac{1}{2}(p_b-\gamma^2b^{-2}p_b)\\

515:  \dot{c} = -2c, &\quad\quad&

516:  \dot{p}_c = 2p_c\, ,

517: \end{eqnarray*}

518: %

519: where the `dot' denotes derivative with respect to the affine

520: parameter of the Hamiltonian vector field. The equations for

521: $(c,p_c)$ can easily be solved by $c(\T)=c_0e^{-2\T}$ and

522: $p_c(\T)=p_c^{(0)}e^{2\T}$. For the other components we obtain

523: $b(\T)=\pm\gamma\sqrt{e^{-(\T-\T_0)}-1}$ for $b$ and, using this

524: result, $p_b(T)=p_b^{(0)}\sqrt{e^{\T+\T_0} -e^{2\T}}$. If we

525: introduce the new time parameter $t:=e^{\T}$ and the constant

526: $m:=\frac{1}{2}e^{\T_0}$, this gives

527: %

528: \begin{eqnarray}

529:  b(t) = \pm\gamma\sqrt{(2m-t)/t}, &\quad\quad&

530:  p_b(t) = p_b^{(0)}\sqrt{t(2m-t)}\\

531:  c(t) = \mp \gamma m p_b^{o}\, t^{-2}\label{3}, &\quad\quad&

532:  p_c(t) = \pm \,t^2\label{4}

533: \end{eqnarray}

534: %

535: where we have fixed the multiplicative constant in the expressions of

536: $p_c(t)$ and $c(t)$, respectively, by requiring that $|p_c|$ be the

537: geometric radius of the 2-sphere orbits of the ${\rm SO(3)}$ symmetry

538: and by using the Hamiltonian constraint. Projections to the $p_b-p_c$

539: plane of typical trajectories are shown in figure \ref{traj2}. It is

540: clear from the figure that $p_c$ can be taken to be the `internal time

541: parameter'. In section \ref{s4.1} this interpretation will let us

542: regard the quantum Hamiltonian constraint as an `evolution equation'

543: with respect to an internal time parameter defined by the eigenvalues

544: of the operator $\hat{p}_c$.

545: 

546: \begin{figure}

547: \begin{center}

548: \includegraphics[width=3in]{traj2.eps}

549: \caption{\footnotesize Dynamical trajectories in the $p_b-p_c$

550: plane. Each trajectory reaches the maximum value $m$ of $p_b$ and

551: meets the $p_c =0$ axis only at the point where $p_b$ also

552: vanishes (i.e., at the `origin'). Solutions related by the

553: reflection-symmetry $p_c \rightarrow -p_c$ define the same metric

554: but carry triads of opposite orientation. For simplicity the

555: `$b$-parity gauge' is fixed by requiring $p_b \ge 0$.}

556: \label{traj2}

557: \end{center}

558: \end{figure}

559: 

560: Note that we have a 2-parameter family of solutions, labeled by

561: $m$ and $p^{(0)}_b$, as one would expect from the fact that the

562: reduced phase space is 2-dimensional. However, in a space-time

563: description we expect only a 1-parameter family, labeled by $m$.

564: This discrepancy can be traced back to the standard tension

565: between the Hamiltonian and space-time notions of gauge (for a

566: detailed discussion in the context of Bianchi models, see

567: \cite{as}, and in spherically symmetric models, \cite{kt,k}). In

568: the Hamiltonian description, rescalings of $p^{(0)}_b$ do not

569: correspond to gauge because they are not generated by any

570: constraint. In the space-time description, on the other hand, they

571: can be absorbed by a rescaling of the coordinate $x$. Therefore,

572: to make contact with the space-time description, let us fix

573: $p^{(0)}_b =1$. Then, the above solution to the evolution

574: equations defines a 1-parameter family of 3-metrics:

575: %

576: \be q_{ab}(t) = \left(\frac{2m}{t} -1\right) \nabla_a x \nabla_b x

577: + t^2 (\nabla_a\vt \nabla_b \vt + \sin^2\vt  \nabla_a

578: \vp\nabla_b\vp ) \ee

579: %

580: defining precisely the Schwarzschild solution of mass $m$,

581: interior to the horizon. It will be useful to note that, in this

582: solution $p_c>0$, $p_b=0$ at the horizon ($t=2m$) and $p_c =0$ and

583: $b,c$ diverge at the singularity.

584: 

585: We will see in the next section that it is convenient to regard

586: the space $\M$ spanned by $p_b$ and $p_c$ as the mini-superspace

587: on which quantum wave functions are defined. (In that discussion,

588: we will not fix the `$b$-parity gauge'; so $p_c$ \emph{and} $p_b$

589: take values on the entire real line.) Let us therefore interpret

590: various regions in $\M$. The triad is non-degenerate everywhere

591: except when $p_b=0$ or $p_c =0$. However, the two degeneracies are

592: of very different geometric and physical origin. Geometrically

593: (i.e., independent of classical dynamics) $p_c=0$ separates two

594: regions with triads of opposite orientations, given by $\sgn\det

595: E=\sgn(p_c p_b^2)$. On the other hand, as (\ref{cotriad})

596: (together with our gauge choice which ensures $a=p_a =0$) shows,

597: the orientation would not have changed even if we had extended

598: $\M$ across $p_b=0$, allowing for negative values for $p_b$. More

599: importantly, while the co-triad (\ref{cotriad}) remains smooth and

600: only becomes degenerate along $p_b=0$, it diverges along $p_c =0$,

601: signaling the classical singularity. This suggests that the line

602: $p_b =0$ corresponds to the horizon and $p_c =0$ to the

603: singularity. This is confirmed from physical considerations

604: involving classical dynamics. Classical solutions can be smoothly

605: extended to the line $p_b=0$ which is reached for $t=2m$. The

606: curvature $B$ remains well-defined there (see (\ref{B})). The

607: other line, $p_c=0$, is very different. Along classical solutions,

608: the line $p_c=0$ is approached as $t\to 0$, whence the components

609: (and invariants) of curvature $B$ also diverge. This line

610: corresponds to the singularity and cannot be crossed by any

611: classical dynamical trajectory. Finally, note that since the line

612: $p_c=0$ separates the two regions of $\M$, where only the triad

613: orientation is reversed, the two get identified in

614: geometrodynamics. Therefore the singularity would lie at a

615: boundary of the mini-superspace in geometrodynamics, making the

616: geometrical meaning of `evolution across the singularity' obscure.

617: Furthermore, since the Arnowitt--Deser--Misner (ADM) variables are

618: based on covariant metrics (rather than contravariant triad

619: densities), as in other homogeneous models \cite{mb6}, one of the

620: ADM variables becomes {\em infinite} at the classical singularity.

621: This further complicates a discussion of the singularity structure

622: and its resolution.

623: 

624: 

625: \section{Quantum Kinematics}

626: \label{s3}

627: 

628: This section is divided into two parts. In the first we introduce

629: the basic quantization procedure and in the second we discuss

630: quantum geometry with emphasis on the classical singularity.

631: 

632: \subsection{Basics}

633: \label{s3.1}

634: 

635: As in the isotropic model \cite{abl}, we will follow the general

636: procedure used in the full theory \cite{alrev,crbook,ttrev}. Thus

637: the elementary configuration variables will be given by holonomies

638: along curves in $M$ and the momenta by fluxes of triads along

639: 2-surfaces in $M$. However, because of symmetry reduction, one

640: need not consider all (piecewise analytic) curves or surfaces; a

641: judiciously chosen, smaller subset suffices to obtain a set of

642: functions that is sufficiently large to separate points of the

643: reduced phase space.

644: 

645: Let us begin with the holonomies. We will restrict ourselves to

646: three sets of curves: those along the $\R$ direction of $M$ with

647: oriented length $\nu L_o$; those along the equator of $\S^2$ with

648: oriented length $\mu$; and those along the longitudes of $\S^2$

649: also with oriented length $\mu$, where all lengths and

650: orientations are defined using the fiducial triad. (Thus $\nu,

651: \mu$ are positive if the tangent to the respective curves are

652: parallel to the triad vectors and negative if they are

653: anti-parallel.) Holonomies along these curves suffice to

654: completely determine any invariant connection (\ref{sym1}).%

655: %

656: \footnote{In the fully gauge fixed setting, we could omit

657: $h_{\vt}$ since it is related to $h_{\vp}$ by conjugation with

658: $\exp\,\tau_3$. However, we will use a more `democratic' approach

659: which is applicable also when $a$ is not made to vanish by gauge

660: fixing.}

661: %

662: \begin{eqnarray}

663:  h^{(\nu)}_x(A) &=& \exp\int_0^{\nu L_o}\md x \tilde{c}\tau_3=

664: \cos\frac{\nu c}{2}+ 2\tau_3\sin\frac{\nu c}{2} \label{hol1}\\

665: h^{(\mu)}_{\vp}(A) &=& \exp-\int_0^{\mu}\md\vp \tilde{b}\tau_1=

666: \cos\frac{\mu b}{2}- 2\tau_1\sin\frac{\mu b}{2} \label{hol2}\\

667: h^{(\mu)}_{\vt}(A) &=& \exp\int_0^{\mu}\md\vt \tilde{b}\tau_2=

668: \cos\frac{\mu b}{2}+ 2\tau_2\sin\frac{\mu b}{2}\, .\label{hol3}

669: \end{eqnarray}

670: %

671: Matrix elements of these holonomies are functions of the reduced

672: connection and constitute our configuration variables. Elements of

673: the algebra they generate are almost periodic functions of $b$ and

674: $c$ of the form $f(b,c) = \sum_{\mu,\nu}\, f_{\mu\nu} \exp

675: {\frac{i}{2}\, (\mu b + \nu c)}$, where $f_{\mu,\nu} \in

676: {\mathbb{C}}$, $\mu,\nu \in \mathbb{R}$ and the sum extends over a

677: finite set. This algebra is the Kantowski-Sachs analog of the

678: algebra of cylindrical functions in the full theory

679: \cite{alrev,crbook,ttrev}. We will therefore denote it by $\cyl$.

680: Consider the $C^\star$ algebra obtained by completing $\cyl$ using

681: the sup norm. The quantum configuration space $\Ab$ is the

682: Gel'fand spectrum of this algebra. From the structure of the

683: algebra, it follows that the spectrum is naturally isomorphic to

684: the Bohr compactification $\bar{\R}_{\rm Bohr}^2$ of the Abelian

685: group $\R^2$ \cite{afw,jw}. (Recall that in the isotropic case

686: \cite{abl}, the quantum configuration space is isomorphic to

687: $\bar{\R}_{\rm Bohr}$.) The Hilbert space $\tilde\H$ is obtained

688: by the Cauchy completion of $\cyl$ with respect to the natural

689: Haar measure $\mu_o$ on the Abelian group $\bar{\R}^2_{\rm Bohr}$;

690: $\tilde\H = L^2(\bar{\R}_{\rm Bohr}^2,\md\mu_o)$. Employing the

691: standard bra-ket notation, we can define a basis $|\mu,\nu\rangle$

692: in $\tilde\H$ via:

693: %

694: \begin{equation} \label{basis}

695:  \langle b,c|\mu,\nu\rangle = e^{\frac{i}{2}\,(\mu b+\nu c)} \qquad

696:  \mu,\nu\in{\R}\, .

697: \end{equation}

698: %

699: This is an orthonormal basis:

700: %

701: \be \langle \mu^\prime, \nu^\prime|\mu,\nu\rangle =

702: \delta_{\mu^\prime\, \mu}\, \delta_{\nu^\prime\, \nu} \,  \ee

703: %

704: where, on the right side, we have the Kronecker symbol, rather

705: than the Dirac delta distribution. Thus, each basis vector is

706: normalizable and has unit norm.

707: 

708: As one would expect, the configuration ---i.e., holonomy---

709: operators $\hat{h}_x^{(\nu)},\, \hat{h}_\vp^{(\mu)},\,

710: \hat{h}_\vt^{(\mu)}$ operate on $\tilde\H$ by multiplication.

711: However, as in the full theory, these operators are not required

712: to be weakly continuous in parameters $\mu$ and $\nu$, whence

713: there are no operators $\hat{b}, \hat{c}$ corresponding to the

714: connection itself. consequently, although we have only a finite

715: number of degrees of freedom, the von-Neumann uniqueness theorem

716: is inapplicable and this quantum theory is inequivalent to a

717: standard `Schr\"odinger quantization' (for a further discussion,

718: see \cite{afw,jw}). We will see that, as in the isotropic model

719: \cite{abl}, this inequivalence has important consequences.

720: 

721: For the momentum operators we consider fluxes of triads along

722: preferred 2-surfaces. Apart from fixed kinematical factors, they

723: are given by components $p_b,p_c$ of triads which, in view of the

724: symplectic structure (\ref{symp}), are represented by operators%

725: %

726: \footnote{In this paper, we use the standard quantum gravity

727: convention, $\lp^2 = G\hbar$. Unfortunately, this is different

728: from the convention $\lp^2 = 8\pi G\hbar$ used in most of the loop

729: quantum cosmology literature. Hence care should be exercised while

730: comparing detailed numerical factors.}

731: %

732: \be \hat{p}_b = -i{\gamma\lp^2}\,\frac{\partial}{\partial b},

733: \quad\quad \hat{p}_c = -2i\gamma\lp^2 \frac{\partial}{\partial c}\,.

734: \ee

735: %

736: %\begin{eqnarray} \label{triadop}

737: % \hat{p}_b &=& -i\frac{\gamma\lp^2}{8\pi}\frac{\partial}{\partial

738: %b}\\

739: % \hat{p}_c &=& -i\frac{\gamma\lp^2}{4\pi}\frac{\partial}{\partial

740: %c}

741: %\end{eqnarray}

742: %

743: Their eigenstates are the basis states (\ref{basis}),

744: %

745: \be \hat{p}_b|\mu,\nu\rangle =

746: \textstyle{\frac{1}{2}}\,\gamma\lp^2\, \mu |\mu,\nu\rangle,\qquad

747: \hat{p}_c|\mu,\nu\rangle = \gamma\lp^2\, \nu

748:  |\mu,\nu\rangle\, . \ee

749: 

750: However, we still have to incorporate the residual gauge freedom

751: which corresponds to a parity reflection in the $b$ degree of

752: freedom. Therefore, only those states in $\tilde\H$ which are

753: invariant under the parity operator $\hat\Pi_b\colon |\mu,\nu\rangle

754: \rightarrow |-\mu,\nu\rangle$ can belong to the kinematical

755: Hilbert space $\H$. A basis in $\H$ is thus given by:

756: %

757: \be \frac{1}{\sqrt{2}}\, [|\mu, \nu\rangle + |-\mu, \nu\rangle

758: ]\ee

759: %

760: Finally, we express the volume operator in terms of triad

761: operators $\hat{p}_b, \hat{p}_c$. Recall that the region of $M$

762: under consideration is of the type ${\cal I} \times \S^2$, where

763: ${\cal I}$ has length $L_o$ with respect to the fiducial metric.

764: From the classical expression (\ref{vol}) of the volume $V$ of

765: this region, it follows that the operator $\hat{V}$ is given by

766: $\hat{V}=4\pi |\hat{p}_b|\, \sqrt{|\hat{p}_c|}$. It is diagonal in

767: our $|\mu, \nu\rangle$ basis and the eigenvalues are:

768: %

769: \begin{equation}

770:  V_{\mu\nu}=2\pi\, \gamma^{3/2}\, |\mu|\, \sqrt{|\nu|}\,\, \lp^3\, .

771: \end{equation}

772: %

773: As in the isotropic case, the volume of our cell ${\cal I} \times

774: \S^2$ with respect to the fiducial metric can be large, but its

775: volume in the `elementary' state $\exp \frac{i}{2}\,(b+c)$, is of

776: Planck size.

777: 

778: 

779: \subsection{Quantum geometry}

780: \label{s3.2}

781: 

782: As in the isotropic case the triad operators $\hat{p}_b$,

783: $\hat{p}_c$ commute whence, in contrast to the full theory, one can

784: construct the triad representation. In this sub-section we will

785: study the quantum Riemannian geometry using this representation.

786: In the next sub-section we will see that the representation is

787: also convenient for dynamics, for it suggests an intuitive

788: interpretation for the action of the Hamiltonian constraint.

789: 

790: Let us then expand a general state in terms of eigenstates of the

791: triad operators, $|\psi\rangle= \sum_{\mu\nu}\psi_{\mu\nu}

792: |\mu,\nu\rangle$, and use the coefficients $\psi_{\mu\nu}$ to

793: represent the state. This wave function $\psi_{\mu\nu}$ is

794: supported on the mini-superspace $\mathcal{M}$ coordinatized by

795: $p_b$ and $p_c$. In the classical theory, already at the kinematical

796: level, the lines $p_b =0$ and $p_c =0$ are special: the co-triad

797: (\ref{cotriad}) becomes degenerate along $p_b=0$ (and the line

798: represents the horizon) while it diverges on the line $p_c=0$ (and

799: this line represents the singularity). It is then natural to

800: examine wave functions $\psi_{\mu\nu}$ which are supported on

801: lines $\mu=0$ and $\nu=0$ and ask, already at the kinematical

802: level, for the nature of quantum geometry they represent.

803: 

804: Let us consider the co-triad  component $\omega_c \equiv

805: \sgn(p_c)\, |p_b|/\sqrt{|p_c|}$ along $\md x$ (see

806: (\ref{cotriad})). Since it involves an inverse power of $p_c$ and

807: since there exist normalizable kets $|\mu, \nu=0\rangle$ of the

808: operator $\hat{p}_c$ with zero eigenvalue, the naive operator

809: obtained by replacing $p_b$ and $p_c$ by corresponding operators

810: is not even densely defined, let alone self-adjoint. To define

811: $\hat\omega_c$, therefore, a new strategy is needed. We will

812: follow what is by now a standard procedure, adapted from

813: Thiemann's analysis in the full theory \cite{tt}. The first step

814: is to express $\omega_c$ in terms of the elementary variables

815: which do have unambiguous quantum analogs ---holonomies and

816: positive powers of $p_b, p_c$--- and Poisson brackets between

817: them. On the classical phase space, we have the exact equality:

818: %

819: \begin{equation} \label{omegaccomm}

820:  \omega_c=\frac{1}{2\pi\gamma

821: G}\tr\left(\tau_3 h_x \{h_x^{-1},\,V\}\right)

822: \end{equation}

823: %

824: %

825: %\begin{equation}

826: % e_c=\frac{1}{2\pi\gamma

827: %G\L}\tr\left(\tau_3h_x^{(\L)}\{h_x^{(\L)-1},V\}\right)

828: %\end{equation}

829: %

830: where $h_x$ is the holonomy along the interval ${\cal I}$, i.e.,

831: along the edge of length $L_o$ (with respect to the fiducial

832: metric) we fixed in our construction of the phase space. Then,

833: replacing $h_x$ and $V$ by their unambiguous quantum analogs and

834: the Poisson bracket by $1/i\hbar$ times the commutator, we obtain

835: %

836: \begin{eqnarray}

837:  \hat{\omega}_c &=& -\frac{i}{2\pi \gamma\lp^2}\, \tr\left(\tau_3

838: \hat{h}_x[\hat{h}_x^{-1},\, \hat{V}]\right)\nonumber\\

839:  &=& -\frac{i}{2\pi \gamma\lp^2}\left(\sin\frac{

840: c}{2}\hat{V}\cos\frac{c}{2}- \cos\frac{c}{2}\hat{V}\sin\frac{

841: c}{2}\right)\,.

842: \end{eqnarray}

843: %

844: %\begin{eqnarray}

845: % \hat{e}_c &=& -\frac{4i}{\gamma\L\lp^2} \tr\left(\tau_3

846: %\hat{h}_x^{(\L)}[\hat{h}_x^{(\L)-1},\hat{V}]\right)\nonumber\\

847: % &=& -\frac{4i}{\gamma\L\lp^2}\left(\sin\frac{\L

848: %c}{2}\hat{V}\cos\frac{\L c}{2}- \cos\frac{\L c}{2}\hat{V}\sin\frac{\L

849: %c}{2}\right)\,.

850: %\end{eqnarray}

851: %

852: It turns out that this operator is diagonalized by our basis

853: $|\mu, \nu\rangle$ of (\ref{basis}),

854: %

855: \begin{equation}

856: \hat{\omega}_c |\mu,\nu\rangle = \frac{1}{4\pi \gamma\lp^2}

857: (V_{\mu,\nu+ 1}-V_{\mu,\nu- 1})\, |\mu,\nu\rangle =

858: \frac{\sqrt{\gamma}}{2}\lp\, |\mu|\, (\sqrt{|\nu+ 1|}-\sqrt{|\nu-

859: 1|})\,\, |\mu,\nu\rangle\,.

860: \end{equation}

861: 

862: 

863: For large $\nu$, far away from the classical singularity, these

864: eigenvalues are very close to the classical expectation

865: $|\omega_c|= |p_b|/\sqrt{|p_c|}$. This is even true for small

866: $\mu$, i.e.\ we do not need to be far from the horizon. Closer to

867: the classical singularity $p_c=0$, on the other hand, the behavior

868: becomes significantly different from the classical one. At the

869: classical singularity itself, where $|p_b|/\sqrt{|p_c|}$ diverges,

870: the eigenvalue of $\hat{\omega}_c$ is zero.%

871: %

872: \footnote{Exact classical identities such as (\ref{omegaccomm})

873: for objects containing inverse powers are available only in

874: homogeneous models. More generally, the bracket between holonomies

875: and volume is more complicated \cite{mb10}. These complications

876: lead to additional correction terms and the resulting operator is

877: no longer bounded on eigenspaces of $\hat{p}_b$.}

878: %

879: 

880: \emph{Remark:} In the above definition of $\hat{\omega}_c$ we

881: needed a holonomy in the $x$-direction. We chose to evaluate it

882: using the interval ${\cal I}$ of length $L_o$ with respect to the

883: fiducial metric. While this choice is natural because this

884: interval appears already in the construction of the phase space,

885: we could have used an interval of length $L_o\delta$ for some

886: $\delta$. Then the resulting operator would also have been bounded

887: on the entire Hilbert space $\tilde\H$,\, and $|\mu, \nu

888: =0\rangle$ would again have been eigenvectors of the operator with

889: zero eigenvalue. Thus, the qualitative properties of the operator

890: are insensitive to $\delta$. Furthermore, from general

891: considerations one can argue \cite{abl} that $\delta$ should be of

892: the order of $1$. However, the value of the upper bound of the

893: spectrum does depend on $\delta$. Therefore, as in the isotropic

894: model, the precise numerical coefficient in this bound should not

895: be attributed physical significance.

896: 

897: Let us now turn to the second triad component $\omega_b = {\rm

898: sgn} p_b\,\, \sqrt{|p_c|}$. Since $\sqrt{|\nu|}$ is a well-defined

899: function of $\nu$ on the entire spectrum of $\hat{p}_c$, we can

900: quantize $\omega_b$ directly. This operator is again diagonal in

901: our basis $|\mu, \nu\rangle$ and its eigenvalues are given by

902: %

903: \begin{equation}

904:  \hat{\omega}_b|\mu,\nu\rangle=

905: \sqrt{\gamma}\lp\sgn(\mu)\sqrt{|\nu|}\,|\mu,\nu\rangle\,.

906: \end{equation}

907: %

908: Since the operator is quantized directly, the eigenvalues are just

909: the ones one would expect from the classical expression

910: (\ref{cotriad}) of $\omega_b$.

911: 

912: The properties of co-triad operators suggest that the lines

913: $\mu=0$ and $\nu=0$ have very different features also in quantum

914: geometry. No quantum effects are manifest at $\mu=0$ which

915: corresponds to the classical horizon. Near and on the line $\nu=0$

916: which represents the singularity, on the other hand, quantum

917: effects are large. In particular, they remove the classical

918: singular behavior of the co-triad. We emphasize, however, that

919: even in homogeneous models, the ultimate test as to whether or not

920: a singularity persists upon quantization can only come from

921: studying the dynamics. The key questions are: Is quantum dynamics

922: well-defined and deterministic across the classical singularity?

923: And, does this come about without significant quantum corrections

924: to geometry near the horizon? Only quantum dynamics will tell us

925: if the indication provided by the properties of the co-triad

926: operators are borne out.

927: 

928: \section{Quantum Dynamics}

929: \label{s4}

930: 

931: This section is divided into three parts. In the first we present

932: the strategy, in the second we obtain the quantum Hamiltonian

933: constraint, and in the third we use it to discuss the consequent

934: resolution of the classical singularity.

935: 

936: \subsection{Strategy} \label{s4.1}

937: 

938: To bring out the similarities and differences between the reduced

939: model and the full theory, it is instructive to begin with the

940: Hamiltonian constraint (\ref{fullham}) of the full theory:

941: %

942: \begin{equation}

943: \ham= \int\md^3 x N e^{-1}\, [\epsilon_{ijk} E^{ai} E^{bj}\,

944: F_{ab}^k- 2(1+\gamma^2) E_i^aE_j^b\, K_{[a}^iK_{b]}^j ] \, .

945: \end{equation}

946: %

947: We begin by noting some simplifications that arise because of

948: spatial homogeneity. Recall that the connection $A$ is related to

949: the spin-connection $\Gamma$ defined by the triad and the

950: extrinsic curvature $K$ via $A= \Gamma + \gamma K$. As remarked in

951: section \ref{s2.1}, symmetries of the model imply that

952:  $\Gamma$ is the standard magnetic monopole

953: connection; it does not depend on the phase space point under

954: consideration. From its expression (\ref{Gamma}), it follows that

955: its curvature $\Omega$ is given by:

956: %

957: \begin{equation}

958: \Omega = -\sin\vt\,\tau_3\, \md\vt \wedge \md\vp

959: \end{equation}

960: %

961: Since it is a `c-number' it can be trivially taken over to quantum

962: theory. Now, the curvature $F$ of the full connection $A$ can be

963: expanded out as:

964: %

965: \be F_{ab} = 2\partial_{[a}A_{b]} + [A_a,\, A_b] = \Omega_{ab} +

966: 2\gamma\partial_{[a}K_{b]} + \gamma^2 [K_a,\, K_b] + \gamma [\Gamma_a,\,

967: K_b]- \gamma [\Gamma_b,\, K_a]\ee

968: %

969: Using the expressions (\ref{Gamma}), (\ref{K}) and (\ref{sym2}) of

970: $\Gamma$, $K$ and $E$, it is straightforward to verify that:

971: %

972: \ba

973:  [\Gamma, \, \gamma K ] = b \tau_2 \cos \vt \md\vt\wedge \md\vp

974: &=& -\gamma \md K\nonumber\\

975: \epsilon_{ijk} (\partial_{[a} K_{b]}^i) \,E^a_j E^b_k &=& 0

976: \label{simplify} \ea

977: %

978: These equations now imply that the integrand of the Hamiltonian

979: constraint can be written simply as%

980: %

981: \footnote{This simplification has a natural origin. Because of

982: (\ref{simplify}), the right hand side is just

983: $\epsilon_{ijk}E^{ai} E^{bj}\, {}^+\!F_{ab}^k$, where ${}^+\!F$ is

984: the curvature of the self-dual connection ${}^+\!A = \Gamma + i

985: K$.}

986: %

987: %

988: \be \ham(x) = e^{-1}\,E^{ai} E^{bj}\, (\epsilon_{ijk} F_{ab}^k-

989: 2(1+\gamma^2) K_{[a}^iK_{b]}^j)  = e^{-1}\,E^{ai} E^{bj}\,

990: (\epsilon_{ijk}\Omega_{ab}^k - 2 K_{[ai} K_{b]j})\, . \ee

991: %

992: Since $\Omega$ is constant on the entire phase space, the

993: non-trivial part of the Hamiltonian constraint is thus contained

994: just in a term quadratic in extrinsic curvature.

995: 

996: To pass to the quantum theory, we have to express the right hand

997: side in terms of holonomies and triads. There are two possible

998: avenues.\\

999: i) Using (\ref{simplify}), the extrinsic curvature terms in the

1000: constraint function $\ham(x)$ can be written in terms of

1001: curvatures $F$ and $\Omega$ so that we have:

1002: %

1003: \be \label{hami}

1004:  \ham(x) = \frac{1}{\gamma^2}\, \epsilon_{ijk}\, e^{-1}\,

1005: E^{ai}E^{bj}\, [(1 +\gamma^2)\, \Omega_{ab}^k - F_{ab}^k]\, .

1006: \ee

1007: %

1008: The idea now is to use the discussion of co-triads in section

1009: \ref{s3.2} to obtain the operator analog of $\epsilon_{ijk} \,

1010: e^{-1}\,E^{ai}E^{bj}$; carry $\Omega$ to quantum theory trivially

1011: and use holonomies to express the field strength $F$ as in the

1012: full theory.\\

1013: ii) Alternatively, since we have completely fixed the gauge

1014: freedom to perform internal $\SU(2)$ rotations, we can regard $K$

1015: itself as a connection. It is a direct analog of the connection

1016: $A$ in the spatially flat cosmologies where $\Gamma$ vanishes.

1017: Therefore, we will denote its curvature by $\F$; $\F = \md K + [K,\,

1018: K]$. Then, the constraint functional can also be expressed as:

1019: %

1020: \be \label{hamii}

1021:  \ham(x) = \frac{1}{\gamma^2}\, \epsilon_{ijk}\,e^{-1}\,

1022: E^{ai}E^{bj}\, [\gamma^2\, \Omega_{ab}^k - \F_{ab}^k] \ee

1023: %

1024: Now, one can again use the discussion of co-triads in section

1025: \ref{s3.2} to obtain the operator analog of $\epsilon_{ijk}\,

1026: e^{-1}\,E^{ai}E^{bj}$; carry $\Omega$ to quantum theory trivially

1027: and use holonomies to express the field strength $\F$. However,

1028: now holonomies have to be constructed using the connection $K$

1029: rather than $A$. Since our Hilbert space $\H$ is built from

1030: holonomies of $A$, at first it seems difficult to express

1031: holonomies of $K$ as operators on $\H$. However, because of

1032: spatial homogeneity, we can calculate $\ham(x)$ at any point. Let

1033: us choose a point on the equator and use holonomies along the

1034: three sets of curves introduced in (\ref{hol1})--(\ref{hol3}).

1035: Using the expressions (\ref{sym1}) and $(\ref{K})$ of $A$ and

1036: $K$, it follows that, along these curves the holonomies of $A$ and

1037: $K$ are equal! Therefore, the required holonomies of $K$ can

1038: indeed be expressed as operators on $\H$.

1039: 

1040: Both these strategies are viable and, since they begin with

1041: \emph{exact} expressions (\ref{hami}) and (\ref{hamii}) of

1042: classical constraints, the difference in the resulting operators

1043: is just a quantization ambiguity. The first strategy is more

1044: natural in the sense that it does not require the introduction of

1045: a new connection. On the other hand, since the new connection $K$

1046: is the direct analog of the connection used in spatially flat

1047: cosmologies (in particular in \cite{abl}), to facilitate

1048: comparisons, the second strategy has been used in general,

1049: homogeneous models \cite{mb6,bdv}. For definiteness, we will use

1050: the second strategy in the main discussion and comment at the end

1051: on how the first strategy modifies the final constraint

1052: operator. The analysis of the resolution of the singularity can be

1053: carried out with either methods and the conclusion is the same.

1054: 

1055: 

1056: \subsection{Quantum Hamiltonian Constraint} \label{s4.2}

1057: 

1058: Starting from the classical Hamiltonian constraint

1059: %

1060: \be \label{ham} \ham = \frac{1}{\gamma^2}\, \int \md^3 x\,\,

1061: \epsilon_{ijk}\, e^{-1}\, E^{ai}E^{bj}\, [\gamma^2\, \Omega_{ab}^k

1062: - \F_{ab}^k]\, ,\ee

1063: %

1064: where the integral is taken over the elementary cell, we wish to

1065: pass to the quantum operator. The overall strategy is the same as

1066: in \cite{abl,mb6,bdv} and the subtleties are discussed in detail

1067: in \cite{abl}. Therefore, here we will present only the main steps

1068: and comment on differences from the isotropic case treated in

1069: \cite{abl}.

1070: 

1071: In the isotropic case, the elementary cell is a cube of length

1072: $L_o$ with respect to the fiducial metric. To express the

1073: curvature of the connection and the co-triad operator in the

1074: Hamiltonian constraint, one uses holonomies along the curves of

1075: variable length $\mu_oL_o$ along the edges of this cell. For each

1076: value of $\mu_o$ one obtains a quantum constraint operator

1077: $\hat{{\cal C}}_{\rm Ham}^{(\mu_o)}$. As discussed in \cite{abl},

1078: considerations from the full theory imply that $\mu_o$ should not

1079: be regarded as a regulator; rather, the $\mu_o$-dependence of this

1080: operator should be regarded as a quantization ambiguity. This

1081: ambiguity could not be fixed within the reduced model itself. But

1082: by making an appeal to results from the full theory, one can fix

1083: the value of $\mu_o$ using the minimum eigenvalue of the area

1084: operator in the full theory. (Recall that area enclosed by a loop

1085: enters while expressing curvature in terms of holonomies.)

1086: 

1087: In the present, Kantowski-Sachs model, the overall procedure is

1088: the same but minor adjustments are necessary because of lack of

1089: isotropy. Since $\S^2$ is compact, our elementary cell $\S^2\times

1090: {\cal I}$ has a geometric edge only along the ${\cal I}$

1091: direction. As in section \ref{s3.1}, we will supplement it with

1092: edges along a longitude and the equator of $\S^2$. Holonomies

1093: along these three edges are then given by

1094: (\ref{hol1})--(\ref{hol3}). Now consider curves of length $L_o\delta$

1095: along ${\cal I}$ and of length $\delta$ each along the equator and the

1096: longitude of $\S^2$. Then, the co-triad function in the

1097: Hamiltonian (\ref{ham}) can be expressed as:

1098: %

1099: \begin{equation}\label{cotraid}

1100: \epsilon_{ijk}\tau^i\, e^{-1}E^{aj}E^{bk}= -2(8\pi\gamma G{\cal

1101: L}_{(k)})^{-1}\, \epsilon^{abc}\;\,\w_c^k\,

1102: h_k^{(\delta)}\{h_k^{(\delta)-1},V\}

1103: \end{equation}

1104: %

1105: where on the right hand side the index $k$ runs over $x,\vt,\vp$

1106: and is summed over; $h_k$ is the holonomy along the edge $k$; and

1107: ${\cal L}_x = L_o\delta$ and ${\cal L}_\vt = {\cal L}_\vp = \delta$. Note

1108: that this is an exact equality for any choice of $\delta$. Components

1109: of the curvature $\F$ can also be expressed in terms of these

1110: holonomies. Set

1111: %

1112: \begin{equation}

1113:  h^{(\delta)}_{ij} = h_i^{(\delta)}h_j^{(\delta)}(h_i^{(\delta)})^{-1}

1114: (h_j^{(\delta)})^{-1}

1115: \end{equation}

1116: %

1117: where $i,j,k$ run over $x,\vt,\vp$. Then, using

1118: (\ref{hol1})--(\ref{hol3}) it is straightforward to verify that

1119: %

1120: \begin{equation} \label{F}

1121: \F_{ab}^i(x)\tau_i=\frac{\w^i_a\w^j_b}{{\cal A}_{(ij)}}\,\,

1122: (h^{{\cal A}_{(ij)}}_{ij}-1)\, +\,\,O((b^2+c^2)^{3/2}\sqrt{\cal

1123: A})

1124: \end{equation}

1125: %

1126: where ${\cal A}_{x\vt}=\delta^2L_o={\cal A}_{x\vp}$ and ${\cal

1127: A}_{\vt\vp}=\delta^2$. For $\F_{x,\vt}$ and $\F_{x,\vp}$, (\ref{F}) is

1128: the standard geometric relation between holonomies around closed

1129: loops, the area they enclose and curvature. For $\F_{\vt,\vp}$, on

1130: the other hand, while (\ref{F}) continues to hold, the standard

1131: geometrical interpretation is no longer available because the

1132: edges of length $\delta$ along the equator and a longitude fail to

1133: form a closed loop. Nonetheless, the standard relation is

1134: meaningfully extended because of spatial homogeneity, $\delta^2$

1135: playing the role of the `effective area' ${\cal A}_{\vt,\vp}$.

1136: 

1137: We now have all the ingredients, $\F$, $\Omega$ and the triad

1138: components, to rewrite the classical constraint (\ref{ham}) in a

1139: way which is suitable for quantization. For the $\F$-term in

1140: (\ref{ham}) we obtain

1141: %

1142: \begin{eqnarray}\label{Fterm}

1143:  &-&\gamma^{-2}\int\md^3x\, \epsilon_{ijk}\, \F_{ab}^i\,\,

1144:  e^{-1}E^{aj}E^{bk}

1145:  = 2\gamma^{-2}\int\md^3x\,\, \tr\left( \F_{ab}^i\tau_i\;

1146: \epsilon_{jkl}\, e^{-1} E^{aj}E^{bk}\tau^l\right)\nonumber\\

1147:  &=& -4(8\pi\gamma^3G)^{-1}\int\md^3x

1148: \det\w\;\tr\left(\epsilon^{ijk}{\cal A}_{(ij)}^{-1}

1149: (h_{ij}^{(\delta)}-1){\cal L}_{(k)}^{-1}h_k^{(\delta)}

1150: \{h_k^{(\delta)-1},V\}\right) \,+O((b^2+c^2)^{3/2}\delta) \nonumber\\

1151: &=& \left[-16\pi(8\pi\gamma^3G\delta^3)^{-1}\sum_{ijk}\epsilon^{ijk}

1152: \tr\left(h_{ij}^{(\delta)}h_k^{(\delta)} \{h_k^{(\delta)},V\}\right)\right]\,

1153: +O\left((b^2+c^2)^{3/2}\delta\right)\, .

1154: \end{eqnarray}

1155: %

1156: Next, using $\Omega = -\sin\vt\md\vt\wedge\md\vp \tau_3$, the

1157: $\Omega$ term in (\ref{ham}) becomes:

1158: %

1159: \begin{eqnarray}\label{Omegaterm}

1160: \int\md^3x\,\epsilon_{ijk}\, \Omega_{ab}^i\,\,

1161: e^{-1}E^{aj}E^{bk}\,\, \; &=& -2\int\md^3x\,

1162: \tr\left(\Omega_{ab}^i\tau_i\;

1163: \epsilon_{jkl}\, e^{-1}E^{al}E^{bk}\tau^l\right)\\

1164: &=& -8(8\pi\gamma G\delta)^{-1}\int\md^3x\sin\vt\;

1165: \tr\left(\tau_3\;{\cal L}_x^{-1}h_x^{(\delta)}\{h_x^{(\delta)-1},V\}\right)\\

1166: &=& -32\pi(8\pi\gamma G\delta)^{-1}\;

1167: \tr\left(\tau_3\;h_x^{(\delta)}\{h_x^{(\delta)-1},V\}\right)\,.

1168: \end{eqnarray}

1169: %

1170: Set

1171: %

1172: \be C^{(\delta)} = \big[-2 (\gamma^3

1173: G\delta^3)^{-1}\,\sum_{ijk}\epsilon^{ijk}

1174: \tr\left(h_{ij}^{(\delta)}h_k^{(\delta)} \{h_k^{(\delta)},V\}\right)\big] -

1175: \big[-4(\gamma G \delta)^{-1}

1176: \tr\left(\tau_3\;h_x^{(\delta)}\{h_x^{(\delta)-1},V\}\right)\, \big]\ee

1177: %

1178: Then the classical Hamiltonian constraint is given by

1179: %

1180: \be \ham = \lim_{\delta \to 0}\, C^{(\delta)}  \ee

1181: %

1182: Since all terms in $C^{\delta}$ are expressed purely in terms of our

1183: elementary variables ---holonomies and triads--- which have direct

1184: operator analogs, passage to quantum theory is now

1185: straightforward. We obtain:

1186: %

1187: \begin{eqnarray}

1188:  \hat{C}^{(\delta)} &=& 2i

1189: (\gamma^3\delta^3\lp^2)^{-1} \tr\left(\sum_{ijk}

1190: \epsilon^{ijk}\hat{h}_i^{(\delta)} \hat{h}_j^{(\delta)} \hat{h}_i^{(\delta)-1}

1191: \hat{h}_j^{(\delta)-1}\hat{h}_k^{(\delta)}

1192: [\hat{h}_k^{(\delta)-1},\hat{V}]+2\gamma^2\delta^2\tau_3 \hat{h}_x^{(\delta)}

1193: [\hat{h}_x^{(\delta)-1},\hat{V}]\right)\nonumber\\\nonumber &=&

1194: 4i(\gamma^3\delta^3\lp^2)^{-1}\left(

1195:  8\sin\frac{\delta b}{2}\cos\frac{\delta b}{2} \sin\frac{\delta

1196: c}{2}\cos\frac{\delta c}{2}

1197:  \left(\sin\frac{\delta b}{2}\hat{V}\cos\frac{\delta b}{2}-

1198:  \cos\frac{\delta b}{2}\hat{V}\sin\frac{\delta b}{2}\right)\right.\\

1199:  && + \left.\left(4\sin^2\frac{\delta b}{2}\cos^2\frac{\delta

1200: b}{2}+\gamma^2\delta^2\right)

1201:  \left(\sin\frac{\delta c}{2}\hat{V}\cos\frac{\delta c}{2}-

1202:  \cos\frac{\delta c}{2}\hat{V}\sin\frac{\delta c}{2}\right)\right)\label{C}

1203: \end{eqnarray}

1204: %

1205: (This is a special case of Eq. (26) in \cite{bdv}.)

1206: 

1207:  The action of this operator on the eigenstates

1208: $|\mu,\nu\rangle$ of $\hat{p}_b$ and $\hat{p}_c$ is given by

1209: %

1210: \begin{eqnarray}

1211:  \hat{C}^{(\delta)} |\mu,\nu\rangle &=&

1212: (2\gamma^3\delta^3\lp^2)^{-1}

1213: \left[ 2(V_{\mu+\delta,\nu}-V_{\mu-\delta,\nu})\right.\\

1214: &&\times(|\mu+2\delta,\nu+2\delta\rangle- |\mu+2\delta,\nu-2\delta\rangle-

1215: |\mu-2\delta,\nu+2\delta\rangle+ |\mu-2\delta,\nu-2\delta\rangle)\nonumber\\

1216: &&+\left.(V_{\mu,\nu+\delta}-V_{\mu,\nu-\delta}) (|\mu+4\delta,\nu\rangle-

1217: 2(1+2\gamma^2\delta^2)|\mu,\nu\rangle+ |\mu-4\delta,\nu\rangle)\right]\,.

1218: \nonumber

1219: \end{eqnarray}

1220: %

1221: 

1222: However, while the classical constraint is a real function on the

1223: kinematical phase space ${\bf \Gamma}$, $\hat{C}^{\delta}$ fails to be

1224: self-adjoint on the kinematical Hilbert space $\H$. We therefore

1225: need to add to it its Hermitian adjoint.%

1226: %

1227: \footnote{Although the self-adjoint form of the Hamiltonian

1228: constraint was discussed briefly in \cite{mb7}, this point was

1229: ignored in the detailed treatment of \cite{abl}. A detailed

1230: treatment of the self-adjoint constraint and its semi-classical

1231: implications can be found in \cite{jw,abw}. It is used in a

1232: crucial way to obtain the physical Hilbert space in \cite{aps}.}

1233: %

1234: The result is an operator $\hat{C}_{\rm grav}^{(\delta)}:=\frac{1}{2}

1235: (\hat{C}^{(\delta)}+\hat{C}^{(\delta)\dagger})$ given by:

1236: %

1237: \begin{eqnarray}

1238:  \hat{C}_{\rm grav}^{(\delta)} |\mu,\nu\rangle &=&

1239: (2 \gamma^3\delta^3\lp^2)^{-1}\,

1240: \left[(V_{\mu+\delta,\nu}-V_{\mu-\delta,\nu}+V_{\mu+3\delta,\nu+2\delta}-

1241: V_{\mu+\delta,\nu+2\delta}) |\mu+2\delta,\nu+2\delta\rangle \right.\nonumber\\

1242: &&- (V_{\mu+\delta,\nu}-V_{\mu-\delta,\nu}+V_{\mu+3\delta,\nu-2\delta}-

1243: V_{\mu+\delta,\nu-2\delta}) |\mu+2\delta,\nu-2\delta\rangle \nonumber\\

1244: &&-(V_{\mu+\delta,\nu}-V_{\mu-\delta,\nu}+V_{\mu-\delta,\nu+2\delta}-

1245: V_{\mu-3\delta,\nu+2\delta}) |\mu-2\delta,\nu+2\delta\rangle \nonumber\\

1246: &&+ (V_{\mu+\delta,\nu}-V_{\mu-\delta,\nu}+V_{\mu-\delta,\nu-2\delta}-

1247: V_{\mu-3\delta,\nu-2\delta}) |\mu-2\delta,\nu-2\delta\rangle \nonumber\\

1248: &&+{\textstyle\frac{1}{2}}(V_{\mu,\nu+\delta}-V_{\mu,\nu-\delta}+V_{\mu+4\delta,\nu+\delta}-

1249: V_{\mu+4\delta,\nu-\delta}) |\mu+4\delta,\nu\rangle \nonumber\\

1250: &&-(1+2\gamma^2\delta^2)(V_{\mu,\nu+\delta}-V_{\mu,\nu-\delta}) |\mu,\nu\rangle

1251: \nonumber\\

1252: &&+\left.{\textstyle\frac{1}{2}}

1253: (V_{\mu,\nu+\delta}-V_{\mu,\nu-\delta}+V_{\mu-4\delta,\nu+\delta}-

1254: V_{\mu-4\delta,\nu-\delta}) |\mu-4\delta,\nu\rangle\right]\,.

1255: \end{eqnarray}

1256: 

1257: Physical states in quantum theory are those which are symmetric

1258: under the `parity operator' $\hat{\Pi}_b$ and lie in the kernel of

1259: the operator $\hat{C}_{\rm grav}^{(\delta)}$. That is, the only

1260: non-trivial quantum Einstein's equations are:

1261: %

1262: \be \label{qee1}  (\Psi| \hat{\Pi}_b = 0 \quad,\quad (\Psi|

1263: \hat{C}_{\rm grav}^{(\delta)} = 0 \ee

1264: %

1265: where $(\Psi|$ is an element of the dual $\cylstar$ of the space

1266: $\cyl$ of finite linear combinations of almost periodic functions

1267: of $b,c$. Let us expand $(\Psi|$ using eigenbras $(\mu,\nu|$ (in

1268: $\cylstar$) of the triad operators:

1269: %

1270: \be (\Psi| = \sum_{\mu,\nu} \, \psi_{\mu,\nu}\, (\mu,\nu| \ee

1271: %

1272: Then, $\psi_{\mu,\nu}$ can be regarded as wave functions in the

1273: triad (or, Riemannian geometry) representation. To exhibit the

1274: action of $\hat{C}_{\rm grav}^{(\delta)}$ on $\psi_{\mu,\nu}$ we are

1275: led to separate the cases $\mu \ge 4\delta$ from $\mu < 4\delta$

1276: because of the absolute-value $|\mu|$ in the volume eigenvalues.

1277: For $\mu \ge 4\delta$, quantum Einstein's equation (\ref{qee1})

1278: becomes

1279: %

1280: \begin{eqnarray} \label{qee2}

1281: \hat{C}_{\rm grav}^{(\delta)}\, \psi_{\mu,\nu}

1282: &=&2\delta(\sqrt{|\nu+2\delta|}+\sqrt{|\nu|})

1283: \left(\psi_{\mu+2\delta,\nu+2\delta}- \psi_{\mu-2\delta,\nu+2\delta}\right)

1284: \nonumber\\

1285: && +(\sqrt{|\nu+\delta|}-\sqrt{|\nu-\delta|})

1286: \left((\mu+2\delta)\psi_{\mu+4\delta,\nu}-

1287: (1+2\gamma^2\delta^2)\mu\psi_{\mu,\nu}+

1288: (\mu-2\delta)\psi_{\mu-4\delta,\nu}\right)\nonumber\\

1289: &&+2\delta(\sqrt{|\nu-2\delta|}+\sqrt{|\nu|})

1290: \left(\psi_{\mu-2\delta,\nu-2\delta}-

1291: \psi_{\mu+2\delta,\nu-2\delta}\right)\nonumber\\

1292: &=& 0\, .

1293: \end{eqnarray}

1294: %

1295: For $\mu=0$, we have

1296: %

1297: \begin{equation} \label{qee3mu0}

1298: \sqrt{|\nu+2\delta|} \psi_{2\delta,\nu+2\delta}

1299: +(\sqrt{|\nu+\delta|}-\sqrt{|\nu-\delta|})

1300:  \psi_{4\delta,\nu}

1301: -\sqrt{|\nu-2\delta|}\psi_{2\delta,\nu-2\delta}= 0\, ;

1302: \end{equation}

1303: %

1304: for $\mu=\delta$:

1305: %

1306: \begin{eqnarray} \label{qee3mu1}

1307: &&2(\sqrt{|\nu+2\delta|}+\sqrt{|\nu|}) \psi_{3\delta,\nu+2\delta}+

1308: 2(\sqrt{|\nu+2\delta|}-\sqrt{|\nu|}) \psi_{\delta,\nu+2\delta}\nonumber\\

1309: && +(\sqrt{|\nu+\delta|}-\sqrt{|\nu-\delta|})

1310: \left(3\psi_{5\delta,\nu}+2\psi_{3\delta,\nu}- (1+2\gamma^2\delta^2)\psi_{\delta,\nu}

1311:  \right)\nonumber\\

1312: &&-2(\sqrt{|\nu-2\delta|}+\sqrt{|\nu|})\psi_{3\delta,\nu-2\delta}-

1313: 2(\sqrt{|\nu-2\delta|}-\sqrt{|\nu|})\psi_{\delta,\nu-2\delta}

1314: = 0\,,

1315: \end{eqnarray}

1316: %

1317: for $\mu=2\delta$:

1318: %

1319: \begin{eqnarray} \label{qee3mu2}

1320: &&(\sqrt{|\nu+2\delta|}+\sqrt{|\nu|}) \psi_{4\delta,\nu+2\delta}-

1321: \sqrt{|\nu|}\psi_{0,\nu+2\delta}\nonumber\\

1322: && +(\sqrt{|\nu+\delta|}-\sqrt{|\nu-\delta|})

1323: \left(2\psi_{6\delta,\nu}+\psi_{2\delta,\nu}- (1+2\gamma^2\delta^2)\psi_{2\delta,\nu}

1324:  \right)\nonumber\\

1325: &&-(\sqrt{|\nu-2\delta|}+\sqrt{|\nu|})\psi_{4\delta,\nu-2\delta}

1326: +\sqrt{|\nu|}\psi_{0,\nu-2\delta}=0\,,

1327: \end{eqnarray}

1328: %

1329: and for $\mu=3\delta$:

1330: %

1331: \begin{eqnarray} \label{qee3mu3}

1332: &&2(\sqrt{|\nu+2\delta|}+\sqrt{|\nu|}) (\psi_{5\delta,\nu+2\delta}-

1333: \psi_{\delta,\nu+2\delta})\nonumber\\

1334: && +(\sqrt{|\nu+\delta|}-\sqrt{|\nu-\delta|})

1335: \left(5\psi_{7\delta,\nu}+2\psi_{\delta,\nu}- 3(1+2\gamma^2\delta^2)\psi_{3\delta,\nu}

1336:  \right)\nonumber\\

1337: &&-2(\sqrt{|\nu-2\delta|}+\sqrt{|\nu|})(\psi_{5\delta,\nu-2\delta}-

1338: \psi_{\delta,\nu-2\delta})=0\,.

1339: \end{eqnarray}

1340: %

1341: For values of $\mu$ not an integer multiple of $\delta$ one can derive

1342: similar expressions, but they will not be used in what follows. As

1343: remarked above, strictly speaking, the role of (\ref{qee2}) is

1344: only to select physical wave functions. However, intuitively it

1345: can also be thought of as the quantum evolution equation.  Recall

1346: that $p_c$ can be regarded as an internal time in the classical

1347: theory of the model. Therefore, in the quantum theory, one may

1348: regard $\nu$ as internal time. Then, (\ref{qee2}) can be

1349: interpreted as the quantum Einstein's equation, which evolves the

1350: wave functions in discrete steps of magnitude $2\delta$ along $\nu$.

1351: 

1352: So far, the parameter $\delta$ ---and hence the size of the

1353: `time-step' --- is arbitrary. In the classical theory, we recover

1354: the Hamiltonian constraint only in the limit $\delta$ goes to

1355: zero. This is because, while the expression (\ref{cotriad}) is

1356: exact for all loops, curvature ${}^oF$  can be expressed in terms

1357: of holonomies only in the limit in which the areas of loops shrink

1358: to zero (see (\ref{F})). In quantum theory, however, the

1359: straightforward limit $\lim_{\delta\to 0} \hat{C}^{(\delta)}$

1360: diverges (for the same reasons as in the isotropic case

1361: \cite{abl}). The viewpoint is that this is occurs because the

1362: limit ignores the quantum nature of geometry, i.e., the fact that

1363: in full quantum geometry, the area operator has a minimum non-zero

1364: eigenvalue. The `correct' quantization of the Hamiltonian

1365: constraint $\ham$ has to take in to account the quantum nature of

1366: geometry.

1367: 

1368: To do so, note first that the holonomies used in the expression of

1369: $\F$ define quantum states with $\mu=\delta$ and $\nu=\delta$. Using these

1370: states, we can calculate the \emph{quantum} area of each of the

1371: three faces of the elementary cell $\S^2\times {\cal I}$, enclosed

1372: by the three curves ---one along $x$, one along a longitude and

1373: one along the equator of $\S^2$. The area operators defined by

1374: faces $S_{x,\vt}, S_{x,\vp}$ and $S_{\vt,\vp}$ are, respectively,

1375: $2\pi |\hat{p}_b|,\, 2\pi |\hat{p}_b|$ and $\pi |\hat{p}_c|$ (see

1376: (\ref{area})). The quantum geometry states defined by the three

1377: holonomies are eigenstates of these area operators. Furthermore,

1378: they have the same eigenvalue: $\pi\gamma\delta \lp^2$. Now, in the

1379: full theory, we know that the area operator has a minimum non-zero

1380: eigenvalue, $a_o = 2\sqrt{3}\pi\gamma \lp^2$. The viewpoint, as in

1381: \cite{abl}, is that in the calculation of the field strength $\F$,

1382: it is \emph{physically} inappropriate to try to use surfaces of

1383: arbitrarily small areas. The best one can do is to shrink the area

1384: of the loop till it attains this quantum minimum $a_o$. This

1385: implies that we should set $\delta = 2\sqrt{3}$. To summarize, by

1386: using input from the full theory, we conclude that the quantum

1387: Hamiltonian constraint is given by $\hat{C}^{(\delta_o)}$ with $\delta_o =

1388: 2\sqrt{3}$.

1389: 

1390: 

1391: \emph{Remarks}:

1392: 

1393: i) Note that the character of the difference equation changes

1394: depending on whether $\mu\ge 4\delta$ or $\mu < 4\delta$. In particular,

1395: for $\mu \ge 4\delta$, the knowledge of the wave function at `times'

1396: $\nu + 2\delta$ and $\nu$ determines only the combination

1397: $\psi_{\mu-2\delta, \nu-2\delta} - \psi_{\mu +2\delta,\nu-2\delta}$ at `time'

1398: $\nu-2\delta$ via (\ref{qee2}). On the other hand, for $\mu =0$ the

1399: wave function at $\nu + 2\delta$ and $\nu$ determines $\psi_{2\delta,

1400: \tau+2\delta}$ completely via (\ref{qee3mu0}). We will return to these

1401: differences in section \ref{s4.3}.

1402: 

1403: ii) As discussed in \cite{abl,mb8,bd} in detail, one can recover

1404: the Wheeler-DeWitt equation with a specific factor ordering by

1405: taking a systematic limit of quantum Einstein's equation. In the

1406: present model, the difference equation reduces to the following

1407: differential equation for the wave function

1408: $\Psi(p_b,p_c)=\psi_{2p_b/\gamma\lp^2, \, p_c/\gamma\lp^2}$ on the

1409: classical minisuperspace $\M$:

1410: %

1411: \be\label{wdw} {16\lp^4}\,\left( \sqrt{p_c}\,\f{\partial^2

1412: \Psi}{\partial p_b \partial p_c} + \f{p_b}{4\sqrt{p_c}}\,

1413: \f{\partial^2 \Psi}{\partial p_b^2}+ \frac{1}{2\sqrt{p_c}}

1414: \frac{\partial\Psi}{\partial p_b} \right) - 4 \f{p_b}{\sqrt{p_c}}

1415: \Psi =0 \ee

1416: %

1417: For $\mu \gg \delta, \nu\gg \delta$, solutions of the discrete equation

1418: (\ref{qee2}) can be approximated by those of this Wheeler DeWitt

1419: equation in a precise sense. In essence, this Wheeler-DeWitt

1420: equation is obtained from (\ref{qee2}) by ignoring quantum

1421: geometry effects, i.e., by an appropriate $\delta \to 0$ limit.  For

1422: the approximation by a differential equation, higher derivatives

1423: of the wave function $\psi$ must be sufficiently small. Thus, the

1424: limit $\delta\to0$ for the equation (\ref{qee2}) exists only under

1425: additional assumptions on the wave function. It does not exist for

1426: the \emph{operator} $\hat{C}_{\rm grav}^{(\delta)}$ by itself.

1427: 

1428: \subsection{Absence of Singularity}

1429: \label{s4.3}

1430: 

1431: Recall that the classical singularity occurs at $p_c =0$. The

1432: Wheeler DeWitt equation (\ref{wdw}) is manifestly singular there.

1433: One can multiply it by $\sqrt{p_c}$ and define a new `internal

1434: time, $\bar{p_c} = \ln p_c$ and obtain a regular equation in

1435: variables $p_b,\bar{p}_c$. However, since $\bar{p}_c \rightarrow

1436: -\infty$ at the singularity, this regular equation does not let us

1437: evolve across the singularity. Thus, the overall situation is the

1438: same as in the isotropic model.

1439: 

1440: To obtain the Wheeler-DeWitt equation from the `fundamental'

1441: difference equation (\ref{qee2}), we had to ignore quantum

1442: geometry effects. Thus, the key question is whether this failure

1443: of the Wheeler-DeWitt equation is an artifact of the approximation

1444: used. Can quantum geometry effects make a qualitative difference

1445: as they do in the isotropic model? We will now argue that the

1446: answer is in the affirmative.

1447: 

1448: Let us analyze the `evolution' given by the difference equation

1449: (\ref{qee2}). Does this evolution stop at $\nu =0$? As in the

1450: isotropic case \cite{mb3,mb7}, we will use the quantum constraint

1451: (\ref{qee2}) as a recurrence relation starting at positive $\nu$

1452: and evolve toward smaller values. However, now a new twist arises

1453: because for generic values $\mu \ge 4\delta$, (\ref{qee2}) determines

1454: only the difference $\psi_{\mu+2\delta,\nu-2\delta}

1455: -\psi_{\mu-2\delta,\nu-2\delta}$ as a function of the initial values of

1456: the wave function. Therefore, quantum Einstein's equation has to

1457: be supplemented with an appropriate boundary condition. From

1458: (\ref{qee2}) which holds for $\mu \ge 4\delta$, one might conclude

1459: that, even if we restrict ourselves to the `lattice' $\mu = n\delta$,

1460: one would have to specify the wave function at $\mu=0, \delta, 2\delta,

1461: 3\delta$ at each `time step'. However, as remarked at the end of

1462: section \ref{s4.2}, the form of the difference equations for $\mu

1463: < 4\delta$ is different. An examination of (\ref{qee3mu0}) and

1464: (\ref{qee3mu1}) reveals that it is in fact sufficient to specify

1465: the wave function $\psi_{\mu,\nu}$ just at $\mu=0$ and $\mu=\delta$ at

1466: each `time-step'. This is a mathematically viable choice of the

1467: boundary condition and heuristically it corresponds to providing

1468: data the `horizon'. It will turn out that the issue of the

1469: resolution of singularity is insensitive to the precise choice of

1470: the boundary condition. We will therefore postpone the discussion

1471: of physically appropriate choices until after the main discussion.

1472: 

1473: Let us then start at some finite positive $\nu$ and carry out a

1474: backward evolution using (\ref{qee2}). In contrast to the

1475: Wheeler-DeWitt equation (\ref{wdw}), the coefficients in the

1476: difference equation (\ref{qee2}) are always regular. At first one

1477: might think that this regularity by itself would be sufficient to

1478: ensure that the evolution would be well-defined across the

1479: singularity. Note however that this need not be the case. For, the

1480: backward evolution continues only as long as the coefficient of

1481: the wave function at $\nu-2\delta$ is non-zero. The key question

1482: therefore is whether this coefficient vanishes. The issue is

1483: subtle. For instance,  the coefficients in the non-self-adjoint

1484: operator $(\hat{C}^{\delta})^\dag$ are also non-singular. However, if

1485: one uses it in place of $\hat{C}^{\delta}_{\rm Ham}$, one finds that

1486: the coefficient vanishes and one can not evolve across the

1487: singularity.

1488: 

1489: For our quantum Einstein's equation (\ref{qee2}), this coefficient

1490: is given by $\sqrt{|\nu-2\delta|}+\sqrt{|\nu|}$. By inspection,

1491: \emph{it never vanishes}. Therefore, the backward quantum

1492: evolution remains well-defined and determines the wave function

1493: not only for $\nu >0$  but also in the new region with $\nu\le 0$.

1494: In this precise sense, the classical black hole singularity can be

1495: traversed using quantum evolution and thus ceases to be a boundary

1496: of space-time.

1497: 

1498: Next, recall from section \ref{s4.1} that we could have begun with

1499: the expression (\ref{hami}) of the Hamiltonian constraint in the

1500: classical theory and then proceeded with quantization. What would

1501: be the status of the singularity resolution with this choice? The

1502: procedure to construct the quantum Hamiltonian constraint would be

1503: identical. However, the final result would be slightly different:

1504: $\gamma^2$ in (\ref{qee2}) would be replaced by $1+\gamma^2$. That

1505: is, in the new quantum Einstein equation, the coefficient of

1506: $\psi_\mu,\nu$ in (\ref{qee2}) would be modified and the equation

1507: would become:

1508: %

1509: \begin{eqnarray} \label{qee4}

1510: &&2\delta(\sqrt{|\nu+2\delta|}+\sqrt{|\nu|}) (\psi_{\mu+2\delta,\nu+2\delta}-

1511: \psi_{\mu-2\delta,\nu+2\delta})\nonumber\\

1512: && +(\sqrt{|\nu+\delta|}-\sqrt{|\nu-\delta|})

1513: \left((\mu+2\delta)\psi_{\mu+4\delta,\nu}- (1+2(\gamma^2+1)

1514: \delta^2)\mu\psi_{\mu,\nu}+

1515: (\mu-2\delta)\psi_{\mu-4\delta,\nu}\right)\nonumber\\

1516: &&+2\delta(\sqrt{|\nu-2\delta|}+\sqrt{|\nu|})(\psi_{\mu-2\delta,\nu-2\delta}-

1517: \psi_{\mu+2\delta,\nu-2\delta})\nonumber\\

1518: &=& 0

1519: \end{eqnarray}

1520: %

1521: (The same replacement holds for $\mu < 4\delta$.) Since the

1522: coefficient of $\psi_{\mu-2\delta,\nu-2\delta}- \psi_{\mu+2\delta,\nu-2\delta}$ is

1523: unaltered, one would again be able to `evolve' across the

1524: singularity. Thus, the conclusion is robust within quantum

1525: geometry.%

1526: %

1527: \footnote{However, the relation between the wave function and the

1528: Wheeler--DeWitt approximation is less direct with this form of the

1529: constraint.}

1530: %

1531: 

1532: To conclude, we will briefly return to the question of the

1533: boundary condition that are needed to make the evolution

1534: well-defined even away from the singularity. As noted earlier, in

1535: the backward evolution given by (\ref{qee3mu0}),\,

1536: $\psi_{2\delta,\nu_o}$ is determined by values of the wave function at

1537: $\nu>\nu_o$, so long as for $\nu_o\not=0$. Furthermore,

1538: (\ref{qee2}) then determines the wave function at $\mu=(2+4n)\delta$

1539: for all $n\in{\mathbb N}$ and at that $\nu_o$. One could now use

1540: pre-classicality arguments \cite{mb4} (i.e.\ require that the wave

1541: function not oscillate on small scales) at large $\mu$ to choose

1542: the boundary values, eliminating entirely the need for specifying

1543: boundary conditions. While this could be reasonable in

1544: semiclassical regimes, for small $|\nu|$ it would be questionable

1545: to use pre-classicality in the $\mu$-direction even if a

1546: pre-classical solution exist (which is not guaranteed, see e.g.\

1547: the analysis in \cite{ck,gd}). More importantly, the slice $\nu=0$

1548: is special because $\psi_{2\delta,0}$ is not determined through

1549: (\ref{qee3mu0}) or otherwise. At this point, the condition

1550: (\ref{qee3mu0}) evaluated at $\nu=-2\delta$ gives instead a condition

1551: for previous values of the wave function, translating into one

1552: condition for the initial values. At $\nu=0$, however,

1553: $\psi_{2\delta,0}$ has to be specified as boundary value in addition

1554: to $\psi_{0,0}$ because, unlike similar situations in isotropic

1555: models, it does not drop out of the evolution: it is needed in the

1556: recurrence (\ref{qee2}) for $\mu=4\delta$, $\nu=2\delta$. There is thus no

1557: reduction in conditions on the wave function from the fact that

1558: $\psi_{2\delta,0}$ drops out of (\ref{qee3mu0}); a condition is simply

1559: transferred from boundary values to initial values. Note however

1560: that, even though the line $\nu=0$ (corresponding to the classical

1561: singularity) does show special behavior, evolution does not break

1562: down there; it is only the boundary value problem which changes

1563: character.

1564: 

1565: While such a boundary value problem is mathematically

1566: well-defined, the resulting theory is not necessarily physically

1567: correct. For, in a physically interesting theory one would expect

1568: that, well away from the singularity, there should exist

1569: semi-classical solutions which are peaked on the classical

1570: trajectories. From Fig.~1 it is clear that the semi-classical

1571: solution peaked on the trajectory labelled by mass $m$ would be

1572: sharply peaked at $\nu= \pm 4m^2$ on the line $\mu=0$, whence

1573: semi-classical states peaked at different classical solutions will

1574: have quite different forms near $\mu=0$. Therefore, the theory

1575: obtained by simply fixing the wave function on (or near) $\mu=0$

1576: is not likely to admit a sufficiently rich semi-classical sector.

1577: Indeed, it is not clear that any boundary condition at (or near)

1578: $\mu=0$ will be physically viable. Rather, one may have to impose

1579: the boundary condition at $\mu=\infty$, e.g., by requiring that

1580: the wave function should vanish there.  Indeed, the form of

1581: dynamical trajectories of Fig.~1 implies that \emph{every}

1582: semi-classical state would share this property. Moreover, the

1583: corresponding condition does hold in the closed isotropic model

1584: with a massless scalar field as source \cite{aps2}. As in that

1585: case, one would expect that these semi-classical states would span

1586: a dense subspace in the physical Hilbert space. If so, requiring

1587: that the wave functions vanish at $\mu=\infty$ would be physically

1588: justified. Heuristics suggest that this strategy is viable for the

1589: Wheeler-DeWitt equation (\ref{wdw}). However, detailed numerical

1590: analysis is necessary to establish that the difference equations

1591: (\ref{qee2}) or (\ref{qee4}) admits solutions satisfying this

1592: condition (or a suitable modification thereof) for all $\nu$ and

1593: that its imposition makes the evolution unambiguous.

1594: 

1595: \section{Discussion}

1596: \label{s5}

1597: 

1598: Results of the last two sections support a general scenario that

1599: has emerged from the analysis of singularities in quantum

1600: cosmology. It suggests that the classical singularity does not

1601: represent a final frontier; the \emph{physical} space-time does

1602: not end there. In the Planck regime, quantum fluctuations do

1603: indeed become so strong that the classical description breaks

1604: down. The space-time continuum of classical general relativity is

1605: replaced by discrete quantum geometry which remains regular during

1606: the transition through what was a classical singularity. Certain

1607: similarities between the Kantowski-Sachs model analyzed here and a

1608: cosmological model which has been studied in detail \cite{aps}

1609: suggest that there would be a quantum bounce to another large

1610: classical region. If this is borne out by detailed numerical

1611: calculations, one would conclude that quantum geometry in the

1612: Planck regime serves as a bridge between two large classical

1613: regions. Space-time may be much larger than general relativity has

1614: had us believe.

1615: 

1616: However, as indicated at the end of section \ref{s4.2} significant

1617: numerical work is still needed before one can be certain that this

1618: scenario is really borne out in the model. Moreover, this is a

1619: highly simplified model. It is important to check if the

1620: qualitative conclusions remain robust as more and more realistic

1621: features are introduced. First, one should extend the analysis so

1622: that the space-time region outside the horizon is also covered. A

1623: second and much more important challenge is incorporation of an

1624: infinite number of degrees of freedom by coupling the model, e.g.,

1625: to a spherically symmetric scalar field. First steps along these

1626: lines have been taken \cite{mbss,hw} and one can see that the

1627: evolution still extends beyond the classical singularity

1628: \cite{mb9}. But a comprehensive treatment still remains a distant

1629: goal.

1630: 

1631: \section*{Acknowledgements:}

1632: 

1633: This work was supported in part by NSF grants PHY-0090091, and

1634: PHY-0354932, the Alexander von Humboldt Foundation, the C.V. Raman

1635: Chair of the Indian Academy of Sciences and the Eberly research

1636: funds of Penn State.

1637: 

1638: \begin{thebibliography}{99}

1639: 

1640: \bibitem{cm} Misner C W 1972  Minisuperspaces, in \textit{Magic

1641: without magic, John A. Wheeler festschrift} ed Klauder J R

1642: (Freeman, San Fransisco)

1643: 

1644: \bibitem{jh-lh} Hartle J B 1995 {Spacetime quantum mechanics and

1645: the quantum mechanics of spacetime}, in \textit{Gravitation and

1646: quantizations: Proceedings of the 1992 Les Houches summer school}

1647: (North Holland, Amsterdam) \textit{Preprint} \texttt{gr-qc/9304006}

1648: 

1649: \bibitem{hh} Hartle J B and Hawking S W 1983 {The wave

1650: function of the universe} \textit{Phys.\ Rev.} \textbf{D28} 2960-2975

1651: 

1652: \bibitem{alrev} Ashtekar A and Lewandowski L 2004 Background

1653: independent quantum gravity: A status report  \textit{Class.

1654: Quant. Grav.} \textbf{21} R53-R152

1655: 

1656: \bibitem{crbook} Rovelli C 2004 \emph{Quantum Ggravity} (Cambridge

1657: University Press, Cambridge)

1658: 

1659: \bibitem{ttrev} Thiemann T 2004 \emph{Modern canonical quantum

1660: general relativity} (Cambridge University Press, Cambridge, at

1661: press) \textit{Preprint} \texttt{gr-qc/0110034}

1662: 

1663: \bibitem{mb3} Bojowald M 2001 Absence of singularity in loop

1664: quantum cosmology \textit{Phys. Rev. Lett.} \textbf{86} 5227--5230

1665: 

1666: \bibitem{mb6} Bojowald M 2003 Homogeneous loop quantum

1667: cosmology \textit{Class.\ Quantum Grav.} {\bf 20} 2595--2615

1668: 

1669: \bibitem{abl} Ashtekar A, Bojowald M and Lewandowski J 2003

1670: \textit{Mathematical structure of loop quantum cosmology} {Adv.

1671: Theor. Math. Phys.} \textbf{7}  233--268

1672: 

1673: \bibitem{aps} Ashtekar A, Pawlowski T and Singh P 2005

1674: Quantum nature of the big bang: An analytical and numerical study

1675: \textit{Preprint} in preparation

1676: 

1677: \bibitem{md} Dafermos M 2003 Black hole formation from a complete

1678: regular past \textit{Preprint} \texttt{gr-qc/0310040}

1679: 

1680: \bibitem{hm} Horowitz G T and Myers R 1995 The value of

1681: singularities \textit{Gen. Rel. Grav.} \textbf{27} 915-919

1682: 

1683: \bibitem{group} Marolf  D 1995 Refined algebraic quantization:

1684: Systems with a single constraint \textit{Preprint}

1685: \texttt{gr-qc/9508015};

1686: Giulini  N and Marolf  D 1999 On the Generality of refined

1687: algebraic quantization \textit{Class. Quant. Grav.} \textbf{16}

1688: 2479--2488; A uniqueness theorem for constraint quantization

1689: \textit{Class. Quant. Grav.} \textbf{16} 2489--2505

1690: 

1691: \bibitem{ab2} Ashtekar A and Bojowald M 2005 {Black hole

1692: evaporation: A paradigm} \textit{Class. Quant. Grav.} \textbf{22}

1693: 3349--3362

1694: 

1695: \bibitem{s'thw} Stevens C R, 't Hooft G and Whiting B F 1994 Black

1696: hole evaporation without information loss, \textit{Class. Quantum.

1697: Grav.} \textbf{11} 621--648

1698: 

1699: \bibitem{sh3} Hayward S 2005 The disinformation problem for black

1700: holes (conference version), \texttt{gr-qc/0504038}

1701: 

1702: \bibitem{m}

1703: Modesto L 2004 The Kantowski-Sachs space-time in loop quantum

1704: gravity \textit{Preprint} \texttt{gr-qc/0411032}

1705: 

1706: \bibitem{kn}

1707: Kobayashi S and Nomizu K 1963 {\em Foundations of differential

1708: geometry} (John Wiley, New York)

1709: 

1710: \bibitem{b}

1711: Brodbeck O 1996 On symmetric gauge fields for arbitrary gauge and

1712: symmetry groups {\em Helv.\ Phys.\ Acta} {\bf 69} 321--324

1713: 

1714: \bibitem{bk}

1715: Bojowald M and Kastrup H A 2000 Symmetry reduction for quantized

1716: diffeomorphism invariant theories of connections {\em Class.\

1717: Quantum Grav.} {\bf 17} 3009--3043

1718: 

1719: \bibitem{as} Ashtekar A and Samuel J 1991 Bianchi cosmologies:

1720: The role of spatial topology \textit{Class. Quant.Grav.} \textbf{

1721: 8} 2191--2215

1722: 

1723: \bibitem{kt}

1724: Kastrup H A and Thiemann T 1994 Spherically symmetric gravity as a

1725: completely integrable system {\em Nucl.\ Phys.\ B} {\bf 425}

1726: 665--686

1727: 

1728: \bibitem{k}

1729: Kucha\v{r} K V 1994 Geometrodynamics of Schwarzschild black holes

1730: {\em Phys.\ Rev.\ D} {\bf 50} 3961--3981

1731: 

1732: \bibitem{tt} Thiemann T 1996 {Anomaly-free formulation of

1733: non-perturbative, four-dimensional Lorentzian quantum gravity}

1734: \textit{Phys. Lett.} \textbf{B380} 257--264;  1998 {Quantum spin

1735: dynamics (QSD)} \textit{Class. Quant. Grav.} \textbf{15}

1736: 839--873;  1998 {QSD V : Quantum gravity as the natural regulator

1737: of matter quantum field theories} \textit{Class. Quant. Grav.}

1738: \textbf{15} 1281--1314

1739: 

1740: \bibitem{afw} Ashtekar A, Fairhurst S and Willis J 2003

1741: Quantum gravity, shadow states, and quantum mechanics

1742: \textit{Class. Quantum Grav.} \textbf{20} 1031--1062

1743: 

1744: \bibitem{jw} Willis J 2004 On the low-energy ramifications and

1745: a mathematical extension of loop quantum gravity \textit{Ph.D.

1746: Dissertation, Penn State}

1747: 

1748: \bibitem{mb1} Bojowald M 2001 {Inverse scale factor in isotropic

1749: quantum geometry} \textit{Phys.\ Rev.} \textbf{D64}

1750: 084018

1751: 

1752: \bibitem{mb10} Bojowald M 2005 Degenerate Configurations,

1753: Singularities and the Non-Abelian Nature of Loop Quantum Gravity

1754: {\em Preprint} \texttt{gr-qc/0508118}

1755: 

1756: \bibitem{bt} Brunnemann J and Thiemann T 2005 Unboundedness of Triad-Like

1757: Operators in Loop Quantum Gravity {\em Preprint} \texttt{gr-qc/0505033}

1758: 

1759: \bibitem{bdv} Bojowald M, Date G and Vandersloot K 2004

1760: {Homogeneous loop quantum cosmology: The role of the spin

1761: connection} \textit{Class. Quantum Grav.} \textbf{21}

1762: 1253--1278

1763: 

1764: \bibitem{bd} Bojowald M and Date G 2004 {Consistency conditions

1765: for fundamentally discrete theories} \textit{Class. Quantum

1766: Grav.} \textbf{21} (2004) 121--143.

1767: 

1768: \bibitem{mb7} Bojowald M 2002 {Isotropic loop quantum cosmology}

1769: \textit{Class. Quantum Grav.} {\bf 19} 2717--2741.

1770: 

1771: \bibitem{abw} Ashtekar A, Bojowald M, Willis J 2005 Modifications

1772: to Friedmann's equation from quantum geometry {\em Preprint} in preparation

1773: 

1774: \bibitem{mb8} Bojowald M 2001 The semiclassical limit of loop

1775: quantum cosmology {\em Class.\ Quantum Grav.} {\bf 18} L109--L116

1776: 

1777: \bibitem{mb4} Bojowald M 2001 Dynamical initial conditions in quantum

1778: cosmology {\em Phys.\ Rev.\ Lett.} {\bf 87} 121301

1779: 

1780: \bibitem{ck}

1781: Cartin D and Khanna G 2005

1782: Absence of pre-classical solutions in Bianchi I loop quantum cosmology

1783: {\em Phys.\ Rev.\ Lett.} {\bf 94} 111302

1784: 

1785: \bibitem{gd}

1786: Date G 2005

1787: Pre-classical solutions of the vacuum Bianchi I loop quantum cosmology

1788: {\em Preprint} \texttt{gr-qc/0505030}

1789: 

1790: \bibitem{aps2} Ashtekar A, Pawlowski T and Singh P 2005

1791: Closed cosmology and quantum geometry {\em Preprint} in preparation

1792: 

1793: \bibitem{mbss} Bojowald M 2004 Spherically symmetric quantum geometry:

1794: states and basic operators \textit{Class.\ Quantum Grav.}

1795: \textbf{21} 3733--3753;

1796: Bojowald M and Swiderski R 2004 The volume operator in spherically

1797: symmetric quantum geometry \textit{Class.\ Quantum Grav.}

1798: \textbf{21} 4881--4900;

1799: Bojowald M and Swiderski R 2005 Spherically symmetric quantum geometry:

1800: Hamiltonian constraint {\em Preprint} in preparation

1801: 

1802: \bibitem{hw}

1803: Husain V and Winkler O 2004 Quantum resolution of black hole

1804: singularities \textit{Preprint} \texttt{gr-qc/0410125}

1805: 

1806: \bibitem{mb9}

1807: Bojowald M 2005 Non-singular black holes and degrees of freedom in

1808: quantum gravity {\em Phys.\ Rev.\ Lett.} {\bf 95} 061301

1809: 

1810: \end{thebibliography}

1811: 

1812: \end{document}

1813: 

1814: 

1815: 

1816: 

1817: non-symmetric constraint

1818: 

1819: \begin{eqnarray}

1820: && (V_{\mu+3\delta,\nu+2\delta}-V_{\mu+\delta,\nu+2\delta}) \psi_{\mu+2\delta,\nu+2\delta}

1821:  -(V_{\mu-\delta,\nu+2\delta}-V_{\mu-3\delta,\nu+2\delta})

1822: \psi_{\mu-2\delta,\nu+2\delta}\nonumber\\

1823: && +\case{1}{2}(V_{\mu+4\delta,\nu+\delta}-V_{\mu+4\delta,\nu-\delta})

1824: \psi_{\mu+4\delta,\nu}

1825:  -(1+\gamma^2\delta^2) (V_{\mu,\nu+\delta}-V_{\mu,\nu-\delta})

1826: \psi_{\mu,\nu}\nonumber\\

1827: &&  \qquad+\case{1}{2}(V_{\mu-4\delta,\nu+\delta}-V_{\mu-4\delta,\nu-\delta})

1828: \psi_{\mu-4\delta,\nu}\nonumber\\

1829: && -(V_{\mu+3\delta,\nu-2\delta}-V_{\mu+\delta,\nu-2\delta})

1830: \psi_{\mu+2\delta,\nu-2\delta}

1831:  +(V_{\mu-\delta,\nu-2\delta}-V_{\mu-3\delta,\nu-2\delta})

1832: \psi_{\mu-2\delta,\nu-2\delta}\nonumber\\

1833: &=& 2\delta\sqrt{|\nu+2\delta|} (\psi_{\mu+2\delta,\nu+2\delta}-

1834: \psi_{\mu-2\delta,\nu+2\delta})\nonumber\\

1835: && +\case{1}{2} (\sqrt{|\nu+\delta|}-\sqrt{|\nu-\delta|})

1836: \left((\mu+4\delta)\psi_{\mu+4\delta,\nu}-

1837: 2(1+\gamma^2\delta^2)\mu\psi_{\mu,\nu}+

1838: (\mu-4\delta)\psi_{\mu-4\delta,\nu}\right)\nonumber\\

1839: &&-2\delta\sqrt{|\nu-2\delta|}(\psi_{\mu+2\delta,\nu-2\delta}-\psi_{\mu-2\delta,\nu-2\delta})\nonumber\\

1840: &=& 0\,. \label{diffeq2}

1841: \end{eqnarray}

1842: 

1843: 

1844: 

1845: 

1846: 

1847: \bibitem{mb2} M. Bojowald, \textit{Loop quantum cosmology IV: Discrete

1848: time evolution} {Class. Quant. Grav.}, {\bf 18} (2001),

1849: 1071--1088.

1850: 

1851: \bibitem{mb3} M. Bojowald, \textit{Absence of singularity in loop quantum

1852: cosmology}, {Phys. Rev. Lett.}, \textbf{86} (2001), 5227--5230.

1853: 

1854: \bibitem{mb4} M. Bojowald, \textit{Dynamical initial conditions in quantum

1855: cosmology}, {Phys. Rev. Lett.}, {\bf 87} (2001), 121301.

1856: 

1857: \bibitem{mb4a} M. Bojowald, \textit{Inflation from quantum geometry},

1858: {Phys. Rev. Lett.}, {\bf 89} (2002), 261301.

1859: 

1860: \bibitem{mb5} M. Bojowald, \textit{Quantization ambiguities in

1861: isotropic quantum geometry}, {Class. Quantum Grav.}, {\bf

1862: 19} (2002), 5113--5130.

1863: 

1864: \bibitem{ai} A. Ashtekar and C. J. Isham,  \textit{Representation

1865: of the holonomy algebras of gravity and non-Abelian gauge

1866: theories}, {Class. Quant. Grav.}, \textbf{9} (1992), 1433--1467.

1867: 

1868: \bibitem{al2} A. Ashtekar and J. Lewandowski, \textit{Representation

1869: theory of analytic holonomy algebras}, in `Knots and Quantum

1870: Gravity', edired by J.C. Baez,  Oxford U.\ Press, Oxford 1994.

1871: 

1872: \bibitem{jb1} J.C. Baez, J.C. \textit{Generalized measures in gauge

1873: theory}, {Lett. Math. Phys.}, \textbf{31} (1994), 213-223.

1874: 

1875: \bibitem{mm} D. Marolf and J. Mour\~ao,  On the support of the

1876: Ashtekar-Lewandowski measure, \textit{Commun. Math. Phys.},

1877: \textbf{170} (1995), 583-606.

1878: 

1879: \bibitem{al3} A. Ashtekar and J. Lewandowski, \textit{

1880: Projective techniques and functional integration}, {Jour. Math.

1881: Phys.}, \textbf{36} (1995), 2170-2191.

1882: 

1883: \bibitem{al4} A. Ashtekar and J. Lewandowski, \textit{Differential

1884: geometry on the space of connections using projective techniques},

1885: {Jour. Geo. \& \ Phys.}, \textbf{17} (1995), 191--230.

1886: 

1887: \bibitem{al8} A. Ashtekar and J. Lewandowski,\textit{Background

1888: independent quantum gravity: A status report} (pre-print, 2003)

1889: 

1890: \bibitem{aa1} A. Ashtekar,  \textit{New variables for classical and

1891: quantum gravity}, {Phys. Rev. Lett.}, \textbf{57} (1986), 2244--2247;\\

1892: \textit{New Hamiltonian formulation of general relativity}, {Phys.

1893: Rev.}, \textbf{D36} (1987), 1587--1602.

1894: 

1895: \bibitem{al5} A. Ashtekar and J. Lewandowski, \textit{Quantum theory of

1896: geometry I: Area operators}, {Class. Quant. Grav.}, \textbf{14}

1897: (1997), A55--A81.

1898: 

1899: \bibitem{almmt} A. Ashtekar, J. Lewandowski, D. Marolf, J. Mour\~ao,

1900: and T. Thiemann, \textit{Quantization of diffeomorphism invariant

1901: theories of connections with local degrees of freedom}, {Jour.

1902: Math. Phys.}, \textbf{36} (1995), 6456--6493.

1903: 

1904: \bibitem{hs} H. Sahlmann, \textit{Some comments on the representation

1905: theory of the algebra underlying loop quantum gravity},

1906: \texttt{gr-qc/0207111}.

1907: \bibitem{lo} J. Lewandowski and A. Okolow, \textit{Diffeomorphism

1908: covariant representations of the holonomy-flux star-algebra},

1909: \texttt{gr-qc/0302059}.

1910: \bibitem{st} H. Sahlmann and T. Thiemann, \textit{On the

1911: superselection theory of the Weyl algebra for diffeomorphism

1912: invariant quantum gauge theories}, \texttt{gr-qc/0302090}.

1913: 

1914: \bibitem{afw} A. Ashtekar, S. Fairhurst and J. Willis,

1915: \textit{Quantum gravity, shadow states, and quantum mechanics},

1916: {Class. Quantum Grav.}, \textbf{20} (2003), 1031-1062.

1917: 

1918: \bibitem{rs} C. Rovelli and L. Smolin, \textit{Spin networks and

1919: quantum gravity}, {Phys. Rev.}, \textbf{D52} (1995), 5743--5759.

1920: 

1921: \bibitem{jb2} J. C. Baez, \textit{Spin networks in non-perturbative

1922: quantum gravity}, in `The Interface of Knots and Physics', edited

1923: by Kauffman, L. (American Mathematical Society, Providence, 1996)

1924: pp. 167--203;\\

1925: \textit{Spin networks in gauge theory}, {Adv. Math.} {\bf 117}

1926: (1996), 253--272.

1927: 

1928: \bibitem{mb7} M. Bojowald, \textit{Isotropic loop quantum cosmology},

1929: {Class. Quantum Grav.}, {\bf 19} (2002), 2717--2741.

1930: 

1931: \bibitem{abl} A. Ashtekar, M. Bojowald and J. Lewandowski,

1932: \textit{Mathematical Structure of Loop Quantum Cosmology},

1933: {Adv. Theor. Math. Phys.}, \textbf{7} (2003) 233--268.

1934: 

1935: \bibitem{abck} A. Ashtekar, J. Baez, A. Corichi and K. Krasnov,

1936: \textit{Quantum geometry and black hole entropy}, {Phys. Rev.

1937: Lett.}, \textbf{80} (1998), 904--907.\\

1938: A. Ashtekar, J. Baez and K. Krasnov, \textit{Quantum geometry of

1939: isolated horizons and black hole entropy}, {Adv. Theo. Math.

1940: Phys.}, \textbf{4} (2000), 1--95.

1941: 

1942: \bibitem{bdv} M. Bojowald, G. Date and K. Vandersloot,

1943: \textit{Homogeneous loop quantum cosmology: The role of the spin

1944: connection}, {Class. Quantum Grav.}, \textbf{21} (2004) 1253--1278.

1945: 

1946: \bibitem{bd} M. Bojowald and G. Date, \textit{Consistency Conditions

1947: for Fundamentally Discrete Theories}, {Class. Quantum Grav.},

1948: \textbf{21} (2004) 121--143.

1949: 

1950: \bibitem{mb8} M. Bojowald, \textit{The semiclassical limit of loop

1951: quantum cosmology}, {Class. Quantum Grav.}, \textbf{18} (2002),

1952: L109--L116.

1953: 

1954: \bibitem{gr} M. Gaul and C. Rovelli, \textit{A generalized Hamiltonian

1955: constraint operator in loop quantum gravity and its simplest

1956: Euclidean Matrix Elements}, {Class. Quantum Grav.}, \textbf{18}

1957: (2001), 1593--1624.

1958: 

1959: \bibitem{area} C. Rovelli and L. Smolin, \textit{Discreteness of area

1960: and volume in quantum gravity}, {Nucl. Phys.} \textbf{B442}

1961: (1995), 593--622; Erratum: {Nucl. Phys.} \textbf{B456} (1995)

1962: 753\\

1963: S. Frittelli, L. Lehner and C. Rovelli (1996) \textit{The complete

1964: spectrum of the area from recoupling theory in loop quantum

1965: gravity}, {Class. Quant. Grav.} \textbf{13}, 2921--2932

1966: 

1967: \bibitem{dm} Marolf, D. \textit{Refined algebraic quantization:

1968: Systems with a single constraint}, \texttt{gr-qc/9508015}.

1969: 

1970: \bibitem{tt1} T. Thiemann, \textit{QSD III: Quantum constraint

1971: algebra and physical scalar product in quantum general

1972: relativity}, {Class. Quantum Grav.} \textbf{15} (1998),

1973: 1207--1247.

1974: 

1975: \bibitem{bk} M. Bojowald, and H. ~A. Kastrup, \textit{Symmetry

1976: reduction for quantized diffeomorphism invariant theories of

1977: connections}, {Class.\ Quantum Grav.}, {\bf 17} (2000),

1978: 3009--3043.

1979: 

1980: \bibitem{mb9} M. Bojowald, \textit{Loop quantum cosmology:

1981: I. Kinematics}, \textit{Class.\ Quantum Grav.}, {\bf 17} (2001),

1982: 1489--1508.

1983: 

1984: %\bibitem{tt2} T. Thiemann, \textit{Kinematical Hilbert spaces for

1985: %fermionic and Higgs quantum field theories}, {Class.\ Quantum

1986: %Grav.}, {\bf 15} (1998), 1487--1512.

1987: 

1988: \bibitem{als} A. Ashtekar, J. Lewandowski and H. Sahlmann,

1989: \textit{Polymer and Fock representations for a Scalar field},

1990: {Class.\ Quantum Grav.} {\bf 20} (2003), L1--11.

1991: 

1992: \bibitem{gp} R. Gambini and J. Pullin, \textit{

1993: Discrete quantum gravity: applications to cosmology},

1994: \texttt{gr-qc/0112033}.

1995: 

1996: \bibitem{ae} A. Erd\'elyi, \textit{Asymptotic

1997: expansions}, Chapter II (Dover, New York, 1956).

1998: 

1999: \bibitem{hm}

2000:  G.~T. Horowitz and J. Maldacena, \textit{The black hole final state},

2001: JHEP {\bf 0402} (2004), 008.

2002: 

2003: \end{thebibliography}

2004: 

2005: 

2006: \end{document}

2007: 

2008: 

2009: 

2010:  \vspace{1cm}

2011: 

2012: OLD:

2013: 

2014: scalars $\phi_1=a\Lambda_1+b\Lambda_2$,

2015: $\phi_2=-b\Lambda_2+a\Lambda_1$, $\phi_3=c\Lambda_3$ yield

2016: configuration space given by point holonomies:

2017: \begin{equation}

2018:  {\cal U}=\{(\exp(A\Lambda_1),\exp(A\Lambda_2),\exp(c\Lambda_3))\}

2019: \end{equation}

2020: 

2021: gauge invariant states are functions $f(A,c)$ even in $A$

2022: 

2023: measure (from triple product of Haar measure):

2024: \begin{equation}

2025:  \int f(A,c)\sin^2\case{A}{2} \sin^2\case{c}{2}\md A\md c

2026: \end{equation}

2027: 

2028: orthonormal basis:

2029: \begin{equation}

2030:  \chi_{j_1,j_2}=\frac{\sin(j_1+\case{1}{2})A}{\sin\case{A}{2}}

2031:  \frac{\sin(j_2+\case{1}{2})c}{\sin\case{c}{2}} \quad,\quad

2032:  \zeta_{j_1,j_2}=\frac{\sin(j_1+\case{1}{2})A}{\sin\case{A}{2}}

2033:  \frac{\cos(j_2+\case{1}{2})c}{\sin\case{c}{2}}

2034: \end{equation}

2035: 

2036: or

2037: 

2038: \begin{equation}

2039:  |n,m\rangle=\frac{\exp(\case{1}{2}imA)\exp(\case{1}{2}inc)}{

2040:   2\sin\case{A}{2}\sin\case{c}{2}}

2041: \end{equation}

2042: 

2043: gauge invariant only $|n,m\rangle-|n,-m\rangle$: without restriction $m\geq1$

2044: 

2045: \subsection*{Basic Operators}

2046: 

2047: \begin{equation}

2048:  \hat{p}_A=\frac{1}{4}\left(\widehat{\Lambda_1^i

2049:  J_{1i}}+\left(\widehat{\Lambda_1^i J_{1i}}\right)^*\right)\quad,\quad

2050:  \hat{p}_c=\frac{1}{2}\left(\widehat{\Lambda_3^i

2051:  J_{3i}}+\left(\widehat{\Lambda_3^i J_{3i}}\right)^*\right)

2052: \end{equation}

2053: 

2054: action can be found by acting on $\tr(h_1^m)\tr(h_3^n)$ and

2055: $\tr(h_1^m)\tr(\Lambda_3h_3^n)$ where $h_1=\exp(A\Lambda_1)$,

2056: $h_3=\exp(c\Lambda_3)$ and symmetrization:

2057: 

2058: \begin{equation}

2059:  \hat{p}_A=-\case{1}{2}i(\case{\md}{\md

2060:  A}+\case{1}{2}\cot\case{A}{2})\quad,\quad \hat{p}_c= -i

2061:  (\case{\md}{\md c}+\case{1}{2}\cot\case{c}{2})

2062: \end{equation}

2063: 

2064: \begin{equation}

2065:  \hat{p}_A|n,m\rangle=\case{1}{4}m|n,m\rangle\quad,\quad

2066:  \hat{p}_c|n,m\rangle=\case{1}{2}n|n,m\rangle

2067: \end{equation}

2068: 

2069: volume operator: $\hat{V}=\sqrt{\Lambda_3^i

2070: J_{3i}J_1^iJ_{1i}}+(\ldots)^*$

2071: 

2072: \begin{equation}

2073:  \hat{V}|n,m\rangle=\case{1}{4}\sqrt{m^2-1}\sqrt{\case{1}{2}|n|}=:

2074:  V_{j(n),j(m)}|n,m\rangle

2075: \end{equation}

2076: 

2077: with $j(k):=(|k|-1)/2$

2078: 

2079: (point) holonomies lead to

2080: 

2081: \begin{equation}

2082:  \cos\case{A}{2}|n,m\rangle=\case{1}{2}(|n,m+1\rangle+|n,m-1\rangle)

2083:  \quad,\quad

2084:  \sin\case{A}{2}|n,m\rangle=\case{1}{2i}(|n,m+1\rangle-|n,m-1\rangle)

2085: \end{equation}

2086: 

2087: and analogously for $\cos\case{c}{2}$, $\sin\case{c}{2}$

2088: 

2089: \subsection*{Inverse Triad Components}

2090: 

2091: \begin{equation}

2092:  m_{IJ}:=q_{IJ}/\sqrt{\det q}=\diag\left(\frac{\sqrt{|p_c|}}{|p_A|},

2093:  \frac{\sqrt{|p_c|}}{|p_A|}, \frac{|p_A|}{|p_c|^{\frac{3}{2}}}\right)

2094: \end{equation}

2095: 

2096: quantization (as in isotropic models) leads to eigenvalues

2097: 

2098: \begin{equation}

2099:  m_{11;n,m}=16\left(\sqrt{V_{j(n),j(m+1)}}-

2100:  \sqrt{V_{j(n),j(m-1)}}\right)^2>0\quad (\mbox{due to }m\geq1)

2101: \end{equation}

2102: 

2103: \begin{equation}

2104:  m_{33;n,m}=16\left(\sqrt{V_{j(n+1),j(m)}}-

2105:  \sqrt{V_{j(n+1),j(m)}}\right)^2 \mbox{ vanishes for }n=0

2106: \end{equation}

2107: 

2108: large $m,|n|$:

2109: $m_{11;n,m}\sim\case{4}{m}\sqrt{|n|/2}\sim\frac{\sqrt{|p_c|}}{p_A}$ and

2110: $m_{33;n,m}\sim\case{m}{4}(|n|/2)^{-\frac{3}{2}}\sim

2111: \frac{p_A}{|p_c|^{\frac{3}{2}}}$

2112: 

2113: quantized co-triad components with eigenvalues:

2114: 

2115: \begin{equation}

2116:  \hat{a}_A:=\sqrt{|\hat{p_c}|}\quad,\quad

2117:  a_{A;n,m}=\sqrt{\case{1}{2}|n|}

2118: \end{equation}

2119: 

2120: \begin{equation}

2121:  \hat{a}_c:=\hat{m}_{33}\hat{p_c}\quad,\quad

2122:  a_{c;n,m}= \sqrt{2} n\sqrt{m^2-1}

2123:  \left(|n+1|^{\frac{1}{4}}-|n-1|^{\frac{1}{4}}\right)^2

2124: \end{equation}

2125: 

2126: (use $\hat{m}_{33}$ since $\hat{p}_c$ not invertible)

2127: 

2128: \subsection*{Comparison with Schwarzschild}

2129: 

2130: $a_A=t$, $|a_c|=\sqrt{(2M-t)/t}$ (mass $M$)

2131: 

2132: horizon at $t=2M\rightarrow a_A>0,a_c=0$: $n\not=0$, $m=1$

2133: 

2134: singularity at $t=0\rightarrow a_A=0,a_c\to\infty$: $n=0$, $a_{c;0,m}$

2135: finite (as with inverse scale factor)

2136: 

2137: \subsection*{Evolution}

2138: 

2139: classical Euclidean constraint: $H^E=-\frac{\sgn

2140: p_c}{\sqrt{|p_c|(p_a^2+p_b^2)}} \left[(a^2+b^2-1)(p_a^2+p_b^2)+

2141: 2cp_c(ap_a+bp_b)\right]$

2142: 

2143: Lorentzian: $H=-\gamma^{-2}\frac{\sgn

2144: p_c}{\sqrt{|p_c|(p_a^2+p_b^2)}} \left[(a^2+b^2+\gamma^2)(p_a^2+p_b^2)+

2145: 2cp_c(ap_a+bp_b)\right]$

2146: 

2147: quantization of Euclidean part:

2148: 

2149: \begin{eqnarray}

2150:  \hat{H}^E &=& \frac{8i}{\gamma\kappa l_P^2}\left[

2151:  8\sin\case{A}{2}\cos\case{A}{2} \sin\case{c}{2}\cos\case{c}{2}

2152:  (\sin\case{A}{2}\hat{V}\cos\case{A}{2}-

2153:  \cos\case{A}{2}\hat{V}\sin\case{A}{2})\right.\\

2154:  && - \left.(1-4\sin^2\case{A}{2}-2\sin^4\case{A}{2})

2155:  (\sin\case{c}{2}\hat{V}\cos\case{c}{2}-

2156:  \cos\case{c}{2}\hat{V}\sin\case{c}{2})\right]\nonumber

2157: \end{eqnarray}

2158: 

2159: action in ``triad representation'' $s_{n,m}$ with $s_{n,-m}=-s_{n,m}$

2160: due to gauge invariance:

2161: 

2162: \begin{eqnarray}

2163:  (\hat{H}^Es)_{n,m} &=& \frac{2}{\gamma

2164:  l_P^2}\left[\case{1}{2}\sqrt{|n+2|/2}

2165:  \left(\case{1}{2}(\sqrt{(m+3)^2-1}-\sqrt{(m+1)^2-1})

2166:  s_{n+2,m+2}\right.\right.\\

2167:  &&\qquad -\left.

2168:  \case{1}{2}(\sqrt{(m-1)^2-1}-\sqrt{(m-3)^2-1}) s_{n+2,m-2}\right)\nonumber\\

2169:  &&- \case{1}{4}(\sqrt{|n+1|/2}-\sqrt{|n-1|/2})(\case{1}{4}\sqrt{(m+4)^2-1}

2170:  s_{n,m+4}- 3\sqrt{(m+2)^2-1}s_{n,m+2}\nonumber\\

2171:  &&\qquad+ \case{7}{2}\sqrt{m^2-1}s_{n,m}- 3\sqrt{(m-2)^2-1}s_{n,m-2}+

2172:  \case{1}{4}\sqrt{(m-4)^2-1} s_{n,m-4})\nonumber\\

2173:  &&- \case{1}{2}\sqrt{|n-2|/2}

2174:  \left(\case{1}{2}(\sqrt{(m+3)^2-1}-\sqrt{(m+1)^2-1})

2175:  s_{n-2,m+2}\right.\nonumber\\

2176:  &&\qquad -

2177:  \left.\left.\case{1}{2}(\sqrt{(m-1)^2-1}-\sqrt{(m-3)^2-1})

2178:  s_{n-2,m-2}\right)\right]\nonumber

2179: \end{eqnarray}

2180: 

2181: using $s_{n,-m}=-s_{n,m}$ if $1\leq m\leq 4$, e.g.\

2182: 

2183: \begin{eqnarray}

2184:  (\hat{H}^Es)_{n,1} &=& \frac{2}{\gamma

2185:  l_P^2}\left[\case{1}{4}\sqrt{3}\sqrt{|n+2|/2}

2186:  (\sqrt{5}s_{n+2,3}-s_{n+2,1})\right.\\

2187:  &&- \case{1}{4}(\sqrt{|n+1|/2}-\sqrt{|n-1|/2})(\case{1}{2}\sqrt{6}

2188:  s_{n,5}- \case{7}{2}\sqrt{2}s_{n,3})\nonumber\\

2189:  &&- \left.\case{1}{4}\sqrt{3}\sqrt{|n-2|/2}

2190:  (\sqrt{5}s_{n-2,3}-s_{n-2,1})\right]\nonumber

2191: \end{eqnarray}

2192: 

2193: and

2194: 

2195: \begin{eqnarray}

2196:  (\hat{H}^Es)_{n,2} &=& \frac{2}{\gamma

2197:  l_P^2}\left[\case{1}{2}(\sqrt{3}-1)\sqrt{|n+2|}

2198:  s_{n+2,4}\right.\\

2199:  &&- \case{1}{4}(\sqrt{|n+1|/2}-\sqrt{|n-1|/2})(\case{1}{4}\sqrt{35}

2200:  s_{n,6}- 3\sqrt{15}s_{n,4}+ \case{13}{4}\sqrt{3} s_{n,2})\nonumber\\

2201:  &&- \left.\case{1}{2}(\sqrt{3}-1)\sqrt{|n-2|}

2202:  s_{n-2,4}\right]\nonumber

2203: \end{eqnarray}

2204: 

2205: Evolution using $n$ as internal discrete time: from $n$ to $n+2$

2206:  determine $\case{1}{2}(\sqrt{(m+3)^2-1}-\sqrt{(m+1)^2-1})

2207:  s_{n+2,m+2}- \case{1}{2}(\sqrt{(m-1)^2-1}-\sqrt{(m-3)^2-1})

2208:  s_{n+2,m-2}$ for all $m$ (modified formula for low $m$), solve

2209:  recursively for $s_{n+2,m}$ by using boundary data $s_{n,1}$ and

2210:  $s_{n,2}$ ``at the horizon'', evolution through $n=0$ (singularity)

2211:  possible as in isotropic models

2212: 

2213: Need boundary data at fixed $n_0$, $n_0+1$, $n_0+2$, $n_0+3$ {\em

2214: and\/} at $m=1,2$, analogous to Wheeler-DeWitt equation: contains

2215: $\frac{\partial}{\partial p_A}\frac{\partial}{\partial p_c}$, so needs

2216: boundary data at ``light cone'' in $(A,c)$-plane

2217: 

2218: %%%%%%%%

2219: 

2220: 

2221: %The latter corresponds to the sign of the transversal

2222: %extrinsic curvature components of $K_{ab}$ since they are obtained

2223: %from the connection and the triad via

2224: %$K_{ab}=(\Gamma_a^i-A_a^i)e_b^i/\gamma$ (later we will see that

2225: %the connection components in this model are proportional to the

2226: %extrinsic curvature components themselves).

2227: 

2228: In what follows we will fix the remaining gauge freedom:

2229: 

2230: %In what follows we will choose the convention that the sign is carried

2231: %by the gauge invariant triad component even though it correponds to

2232: %the sign of extrinsic curvature. This has the advantage that the gauge

2233: %invariant connection component

2234: %\begin{equation}

2235: % A:=\sqrt{a^2+b^2}

2236: %\end{equation}

2237: %which by definition is always positive, does not depend on the triad

2238: %components. For the gauge invariant triad component we then have

2239: %\begin{equation}

2240: % p_A:=\frac{ap_a+bp_b}{\sqrt{a^2+b^2}}

2241: %\end{equation}

2242: %which can be positive or negative. It is important to keep in mind,

2243: %however, that the sign of $p_A$ does not influence the orientation of

2244: %the triad since that would be given by

2245: %$\sgn(p_cp_A^2)=\sgn(p_c)$. Instead, the sign information of $p_A$

2246: %corresponds to the sign of extrinsic curvature, as explained

2247: %above. (When viewing the system as a cosmological Kantowski--Sachs

2248: %model, putting the sign in the triad components has the additional

2249: %advantage that the recollapse can be studied in the triad

2250: %representation.)

2251: 

2252: %ALTERNATIVELY: Thus, it is natural to put the sign

2253: %into the gauge invariant connection component by defining

2254: %\begin{equation}

2255: % A:=\sgn(ap_a+bp_b) \sqrt{a^2+b^2}\,.

2256: %\end{equation}

2257: %For the triad, on the other hand, we only need one sign factor in

2258: %order to indicate the orientation which is given by $\sgn(p_c)$. Thus,

2259: %we do not loose any information if be use the gauge invariant triad

2260: %component

2261: %\begin{equation}

2262: % p_A:=\sqrt{p_a^2+p_b^2}

2263: %\end{equation}

2264: %which by definition is always positive. In fact, the orientation would

2265: %be given by $\sgn(p_cp_A^2)$ which would not distinguish between

2266: %positive and negative $p_A$.

2267: 

2268: 

2269: 

2270: \end{document}

2271: