1: %\documentclass[preprint,aps,preprintnumbers,amsmath,amssymb]{revtex4}

2: \documentclass[preprint,nofootinbib,aps,preprintnumbers,amsmath,amssymb]{revtex4}

3: %documentclass[11pt]{JHEP3}

4: %\documentclass[12pt]{JHEP3}

5: \usepackage{graphicx}% Include figure files

6: \usepackage{dcolumn}% Align table columns on decimal point

7: \usepackage{bm}% bold math

8: \usepackage{natbib}

9: 

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

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

12: \newcommand{\beq}{\begin{equation}}

13: \newcommand{\eq}{\end{equation}}

14: \newcommand{\bea}{\begin{eqnarray}\displaystyle}

15: \newcommand{\eea}{\end{eqnarray}}

16: \newcommand{\ea}{\end{eqnarray}}

17: \newcommand{\ph}{$\phi$ }

18: \newcommand{\ch}{$\chi$ }

19: \newcommand{\p}{\partial}

20: \newcommand{\s}{\sigma}

21: \newcommand{\al}{\alpha}

22: \newcommand{\half}{\frac{1}{2}}

23: \newcommand{\pp}{p^+}

24: \newcommand{\vol}{V_0^{(7)}}

25: 

26: %%  Adjust these for your printer:

27: \textwidth=6.3in  \textheight=8.2in

28: \oddsidemargin=0.1in   \topmargin=.03cm

29: %%%%%%%%%%%%%%%%%%%%%%%% BEGIN DOCUMENT %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

30: \begin{document}

31: 

32: \title{Dilaton Dynamics from Production of Tensionless Membranes}

33: 

34: \author{Sera Cremonini}

35: \email{sera@het.brown.edu}

36: \affiliation{Physics Department, Brown University\\

37: Providence, RI 02906}

38: \author{Scott Watson}

39: \email{watsongs@physics.utoronto.ca}

40: \affiliation{Physics Department, University of Toronto \\

41: Toronto, ON}

42: 

43: \date{\today}

44: 

45: \begin{abstract}

46: In this paper we consider classical and quantum corrections to

47: cosmological solutions of $11$D SUGRA coming from dynamics of

48: membrane states. We first consider the supermembrane spectrum

49: following the approach of Russo and Tseytlin for consistent

50: quantization. We calculate the production rate of BPS membrane

51: bound states in a cosmological background and find that such

52: effects are generically suppressed by the Planck scale, as

53: expected. However, for a modified brane spectrum possessing

54: enhanced symmetry, production can be finite and significant. We

55: stress that this effect could not be anticipated given only a

56: knowledge of the low-energy effective theory. Once on-shell,

57: inclusion of these states leads to an attractive force pulling the

58: dilaton towards a fixed point of S-duality, namely $g_s=1$.

59: Although the SUGRA description breaks down in this regime,

60: inclusion of the enhanced states suggests that the center of

61: M-theory moduli space is a dynamical attractor. Morever, our

62: results seem to suggest that string dynamics does indeed favor a

63: vacuum near fixed points of duality.

64: \end{abstract}

65: 

66: \maketitle

67: \tableofcontents

68: \newpage

69: 

70: \section{Introduction}

71: Low-energy descriptions of string theory generically predict the

72: existence of a large number of scalar fields,

73: or moduli, which are associated with the size and shape of the extra

74: dimensions,

75: as well as the position and orientation of any branes present in the

76: theory.

77: These moduli are of interest for a number of reasons.

78: From a theoretical point of view, the different vacuum expectation

79: values (VEVs) of these

80: fields correspond to different choices for the string vacuum, leading

81: to an indeterminacy of the theory.

82: From a more phenomenological viewpoint, light scalars can have a profound

83: effect

84: on both the early and late universe. If not fixed, these moduli can

85: lead to a period of

86: inflation, modifications of fundamental constants, and violations of

87: the equivalence principle.

88: If fixed, one must worry about the mass scales involved and the

89: effects of the resulting relic density on cosmological

90: observations, e.g. on the cosmic microwave background

91: or Big Bang nucleosynthesis.

92: 

93: Recently it has been argued \cite{Vafa:2005ui} (see also

94: \cite{Arkani-Hamed:2006dz}) that, by taking low-energy

95: supergravity (SUGRA) as the effective description for string theory, we

96: may

97: miss certain crucial aspects of the underlying theory.

98: As an example, when attempting to understand the vacuum

99: structure of the theory one

100: should remember that moduli space must be of finite size in order to

101: have a realistic theory of gravity (finite $G_N$).

102: Other crucial aspects that we wish to address in this paper are the

103: role of dualities and the importance of dynamics.

104: That is, as background fields evolve the effective

105: mass of heavy states that are outside the realm of the low-energy

106: theory can change.

107: In fact, near points of enhanced symmetry (ESPs), frequently associated

108: with dualities, these additional states can become massless and

109: therefore play a vital role in the low-energy theory.

110: It is important that these states lie beyond the naive low-energy

111: SUGRA description and, without an underlying knowledge of the

112: fundamental theory,

113: effects associated with these additional degrees of freedom (such as

114: particle creation and radiative corrections) would be missed.

115: Indeed, recent work suggests that including these effects can have

116: important and interesting effects both in string theory and cosmology

117: \cite{Silverstein:2003hf,Kofman:2004yc,Watson:2004aq,Alishahiha:2004eh}.

118: 

119: In this paper we will consider cosmological solutions of $11$D SUGRA

120: and ask what corrections, if any, result from the dynamics of the

121: moduli (i.e., dilaton and radii).  Focusing on the case of particle

122: production, we show that production of BPS membrane bound states,

123: whose mass depends on the evolution of the moduli, is generally

124: suppressed by the Planck scale.  This is expected and assures us of

125: the validity of the low-energy effective theory.  However, in the

126: special case of membranes that become tensionless near fixed points of

127: duality, we find that significant production can occur.  We are unaware of an explicit construction of such states, however we anticipate their existence given ubiquitous examples in the

128: lower-dimensional string theory case.

129: 

130: We find that by including enhanced states in the low-energy theory, the

131: center of the M-theory moduli space becomes an attractor, and the

132: evolution of the moduli tends toward fixed points of duality.  In the

133: case of the dilaton, which can be taken to be $R_{11}$ in the $11$D

134: theory, we find that the evolution leads to the region near $g_s=1$,

135: i.e. strong coupling.

136: Naively trusting the

137: equations of motion in this region, we find that all moduli will in

138: fact become trapped and we recover a three dimensional radiation

139: dominated universe at late times. Although we expect additional

140: corrections to arise near the regime of strong coupling, it is

141: intriguing that our results seem to provide an

142: explicit example of the belief (\cite{Brustein:2002xf} and

143: \cite{Dine:1998qr}) that a string vacuum consistent with gravity and

144: particle phenomenology should lie near fixed points of duailty.

145: 

146: 

147: The paper is organized as follows.  In Section {\ref{MthStates} we

148: review a scheme developed by Russo and Tseytlin for quantizing the

149: supermembrane in special limits.  This will allow us to obtain the BPS

150: spectrum of membrane bound states, which we will utilize throughout

151: the paper.

152: In Section \ref{ClTrapping} we briefly discuss cosmology with these

153: states and show that they can lead to stabilization, but we also point out

154: that

155: such considerations seem inconsistent, since they lie outside the scope

156: of the effective theory.  In Section \ref{QmTrapping} we consider the

157: possibility of producing membrane bound states within a cosmological

158: background of $11$D SUGRA.  After reviewing the basic formalism for

159: calculating production, we find an expression for the production rate

160: and find that membrane bound states suffer exponential suppression, as

161: expected.  However, we also find that, for a special class of states

162: possessing enhanced symmetry, significant production can result.

163: In Section \ref{cosmo} we consider the modified effective theory in

164: the presence of the produced states, and determine the new evolution

165: of the moduli.  We find an attractor behavior towards the center of

166: moduli space, and trusting the equations in the center region leads to

167: trapping of the moduli.  We conclude with a discussion of the limitations of our approach and future work in progress.

168: 

169: \section{M-theory and the Supermembrane \label{MthStates}}

170: 

171: In this section we review progress that has been made in understanding M-theory

172: as a fundamental theory of membranes.

173: The goal will be to gain an understanding of the spectrum of the fundamental theory

174: beyond the low-energy effective description of $11$D SUGRA.

175: We will be concerned with finding membrane states whose mass depends explicitly

176: on the low-energy moduli, namely the dilaton and radii of the extra dimensions.

177: 

178: We begin by reviewing the work of \cite{Russo:1996if}, where

179: the authors were able to clarify the $11$D origin of many

180: lower-dimensional solutions, as well as identify states of the supermembrane

181: with known string theory configurations.

182: Much effort has been devoted to understanding composite BPS

183: configurations of branes in $11$D.

184: We will concentrate in particular on non-threshold (non-zero

185: binding energy) bound states, the canonical example of which (in string theory) is

186: the $(p,q)$ string, a bound

187: state of a NS-NS string and a R-R string in type IIB theory.

188: 

189: In \cite{Schwarz:1995du,Schwarz:1996bh}, it was

190: suggested that the $(p,q)$ string states should be related to the

191: BPS states of a wrapped M$2$-brane in $11$D. This was supported by a

192: comparison of the {\it zero-mode} contributions of the membrane to the

193: string mass spectra, where agreement was found. However, matching at

194: the oscillator level was not shown, due to the intrinsic

195: difficulties associated with the quantization of the supermembrane.

196: This difficulty was overcome and

197: the {\it oscillating} contributions to the spectrum were

198: obtained in \cite{Russo:1996if}, where supermembrane

199: quantization was achieved in a specific limit in which the theory

200: dramatically simplifies.

201: 

202: Some of the M-theory

203: solutions studied in \cite{Russo:1996if} can have interesting

204: cosmological consequences, as will become apparent in the next section.

205: Specifically, we are interested in certain M-theory states

206: which have a classical supergravity description in terms of

207: a bound state of an M$2$-brane wrapping a $T^2$

208: and of a gravitational wave propagating along one of the torus directions.

209: We will start by presenting the 11D SUGRA solution for such configurations, and show

210: that in a certain limit it can be reduced to the type IIB $(p,q)$ string.

211: Since we are primarily interested in obtaining the mass spectrum of these states, we will

212: then present their interpretation in supermembrane theory.

213: Specifically, we will outline the derivation of the spectrum of such

214: solutions as given in \cite{Russo:1996if}, where it was first confirmed that

215: the mass of an oscillating membrane does indeed agree with that of the $(p,q)$ string.

216: Finally, at the end of this section we attempt to motivate a new class of states possessing enhanced symmetry.  We will save an explicit construction of such states for future work, here only comparing their spectrum to the know BPS states of the supermembrane.  We will see in future sections that such states are the only possibility for membrane production in a low-energy theory.

217: 

218: \subsection{D=11 SUGRA Solutions}

219: 

220: We present a concise introduction to the 11D origin of certain

221: non-marginal BPS configurations of type II string theories.

222: We consider the 11-dimensional space $M^9 \times T^2$.

223: The torus coordinates are labeled by $(y_1,y_2)$ and have periods

224: $(2\pi R_9,2\pi R_{11})$, while the spatial coordinates of $M^9$ are denoted by $x_i$, $i=1,\ldots,8$.

225: We are ultimately interested in considering a bound state of an M$2$ brane,

226: which wraps the $T^2$, and of a gravitational wave propagating in

227: the $y_2$ direction. We will briefly show how to construct such states

228: from simpler solutions.

229: 

230: The 11D solution representing a gravitational wave moving along $y_1$ is given by

231: \bea

232: ds^2 &=& -dt^2 + dy_1^2 + dy_2^2 + W (d t+dy_1)^2+ dx_i dx_i \, ,\nonumber \\

233: C_{\mu \nu \rho} &=& 0, \; \; \; W=\frac{\tilde{Q}}{r^6}, \; \; \;

234: r^2=x_i x_i.

235: \ea

236: Since $y_1$ is periodic, the charge $\tilde{Q}$ must be quantized

237: in units of $1/R_9$; with the correct normalization factor, the

238: charge becomes $\tilde{Q}=c_0 \frac{l}{R_9}$.

239: This solution is found to preserve $1/2$ of the supersymmetries.

240: The other basic 11D object that we will need is the 2-brane, with

241: solution \cite{Russo:1996if},

242: \bea

243: ds^2 &=& H_2^{-\frac{2}{3}} \, \left[-dt^2 + dy_1^2 + dy_2^2 \right] + H_2^{\frac{1}{3}} dx_i

244: dx_i \, , \nonumber \\

245: C_3 &=& H_2^{-1} dt \wedge dy_1 \wedge dy_2, \;\;\;

246: H_2=1+\frac{Q}{r^6}.

247: \ea

248: The charge in this case is given by $Q=c_0 \frac{w_0 R_9}{\alpha'}$, where

249: $w_0$ denotes winding around the target torus (we will define

250: $w_0$ more precisely in the next section).

251: 

252: 

253: Combining the 2-brane solution with a gravitational wave moving in

254: an {\it arbitrary} direction gives

255: \bea

256: ds^2 &=& H_2^{-\frac{2}{3}} \, \bigl[-dt^2 + dy_1^2 + dy_2^2 + W (d t-dz_1)^2\bigr]+ H_2^{\frac{1}{3}} dx_i

257: dx_i\, , \nonumber \\

258: C_3 &=& H_2^{-1} dt \wedge dy_1 \wedge dy_2, \; \; \; z_1=y_1

259: \cos{\theta}+y_2\sin{\theta}\, ,

260: \ea

261: where the gravitational wave charge is modified

262: in the following way:

263: \bea

264: W&=&\frac{\tilde{Q}_q}{r^6}, \; \; \; \tilde{Q}_q=c_0\sqrt{\frac{l_9^2}{R_9^2}+\frac{l_{11}^2}{R_{11}^2}}

265: =\tilde{Q}\sqrt{q_1^2+q_2^2 \tau_2^2},\\

266: \text{with}&& \; \; \; \cos{\theta}=\frac{q_1}{\sqrt{q_1^2+q_2^2 \tau_2^2}}, \; \; \; \; \;

267: \tau_2=\frac{R_9}{R_{11}}.

268: \ea

269: After appropriate dimensional reductions and applications of dualities,

270: this solution can be identified with a boosted BPS bound state of

271: an $F1$ string and $D0$ brane in Type IIA string theory, and with

272: a boosted $(p,q)$ string in Type IIB.

273: 

274: We are specifically interested in the 2-brane + wave solution with

275: a boost along $y_2$ only (with $\cos{\theta}=0$  and $\sin{\theta}=1$),

276: giving \cite{Russo:1996if},

277: \bea

278: ds^2 &=& H_2^{-\frac{2}{3}} \, \bigl[-dt^2 + dy_1^2 + dy_2^2 + W (d t-dy_2)^2\bigr]+ H_2^{\frac{1}{3}} dx_i

279: dx_i, \nonumber \\

280: C_3 &=& H_2^{-1} dt \wedge dy_1 \wedge dy_2,

281: \ea

282: where

283: \bea

284: H_2&=&1+\frac{Q}{r^6}, \; \; \; r^2=x_i x_i,  \; \; \; Q= (4\pi^2 T_3 c_0 w_0 R_9)\, R_{11}, \\

285: W&=& \frac{\tilde{Q}}{r^6},  \; \; \; \tilde{Q}=c_0

286: \frac{l_{11}}{R_{11}},

287: \ea

288: and we used the definition of $\alpha'$. This state is BPS, preserving $1/4$ of the supersymmetries.

289: 

290: To obtain the mass spectrum of the M-theory membrane that

291: corresponds to this classical supergravity solution, we will make use of

292: supermembrane theory.

293: Before proceeding, we would like to stress the advantage of studying

294: BPS solutions. The BPS condition guarantees that the above classical

295: SUGRA solutions will exhibit some of the features of the full

296: quantum (string) theory, since these states are protected from quantum corrections.

297: That is, if we construct these states in a limit of the theory that is well understood, we can then

298: extrapolate them to regimes that are less understood,

299: giving us a partial knowledge of the spectrum of the theory.

300: 

301: \subsection{Supermembrane Mass Spectrum}

302: 

303: The original studies of the physical spectrum of wrapped membranes of toroidal topology

304: have been performed in \cite{Bergshoeff:1987qx,Duff:1987cs}.

305: Quantization of the supermembrane is highly non-trivial, and one

306: might even wonder whether a consistent quantum theory of the

307: supermembrane can be defined. Addressing the {\it oscillating}

308: membrane is particularly difficult, since it involves dealing with

309: a highly non-linear interacting theory.

310: In the work of \cite{Russo:1996if}, however, quantization was

311: achieved in an appropriate limit in which the interacting terms

312: dropped out, and the theory could be solved.

313: Next, we will outline the arguments of

314: \cite{Russo:1996if} leading to the mass spectrum of the oscillating

315: membrane.

316: 

317: We consider a membrane on $\mathcal{M}^9 \times T^2$. The compact

318: directions are labelled by $X^9$ and $X^{11}$, with radii $R_9,

319: R_{11}$.

320: Wrapping the membrane around the toroidal directions

321: gives

322: \beq

323: \label{X1X2}

324: X^9(\sigma,\rho)= w_9 \, R_9\, \sigma +

325: \tilde{X}^9(\sigma,\rho)\, , \;\;\;\;\;\;\;\;\;

326: X^{11}(\sigma,\rho)= w_{11} \, R_{11}\, \rho +

327: \tilde{X}^{11}(\sigma,\rho)\, ,

328: \eq

329: where the (single-valued) functions

330: $\tilde{X}^9(\sigma,\rho)$ and $\tilde{X}^{11}(\sigma,\rho)$

331: can be expanded in a complete basis of functions on the torus:

332: \beq

333: \label{X1X2tilde}

334: \tilde{X}^9(\sigma,\rho)=\sqrt{\alpha'}\,\sum_{k,m}X^9_{(k,m)}\,

335: e^{ik\sigma+im\rho} , \;\;\;\;\;\;

336: \tilde{X}^{11}(\sigma,\rho)=\sqrt{\alpha'}\,\sum_{k,m}X^{11}_{(k,m)}\,

337: e^{ik\sigma+im\rho}\, . \eq The constants $\alpha'$ and $T_3$ are

338: defined as \beq \alpha'=\frac{1}{4 \pi^2 R_{11} T_3}, \; \; \; \;

339: \; T_3=\frac{l_P^{-3}}{4\pi^2}.

340: \eq

341: It will turn out to be convenient to define a

342: winding number $w_0$ in the following way,

343: \beq

344: w_0 =

345: \frac{1}{(2\pi R_9) (2\pi R_{11})} \int d\sigma d\rho \,

346: \{X^9,X^{11}\}=w_9 w_{11}, \;\;\;\;\; \{X,Y\}\equiv\p_\sigma X\p_\rho

347: Y-\p_\rho X\p_\sigma Y, \nonumber

348: \eq

349: counting how many times the

350: membrane is wound around the target torus. A membrane with $w_0

351: \neq 0$ is stable for topological reasons \cite{Russo:1996ph}, as

352: will become explicit in the Hamiltonian description. This is a

353: motivation for wrapping the membrane on $\mathcal{M}^9 \times T^2$,

354: and not on $\mathcal{M}^{10} \times S^1$, which would give $w_0=0$.

355: 

356: The transverse (single-valued) coordinates $X^i$, $i=1, \ldots ,8$, and their

357: corresponding canonical momenta can also be expanded on the torus:

358: \beq

359: \label{Xi}

360: X^i(\sigma,\rho)=\sqrt{\alpha'}\,\sum_{k,m}X^i_{(k,m)}\,

361: e^{ik\sigma+im\rho}, \; \; \; \; \; \; \; \;

362: P^i(\sigma,\rho)=\frac{1}{4 \pi^2

363: \sqrt{\alpha'}}\,\sum_{k,m}P^i_{(k,m)}\,

364: e^{ik\sigma+im\rho}.

365: \eq

366: The (bosonic) light-cone Hamiltonian for the

367: supermembrane is given by

368: \beq

369: H_B=2\pi^2 \int d\s d\rho \bigl[(P^{\,c})^{2} + \frac{1}{2}T_3^2 (\{X^c,X^d\})^2

370: \bigr],

371: \eq

372: where $c,d=1,\ldots,10$.

373: Here we neglect the fermionic sector, which can however be

374: incorporated.

375: Making use of the expansions (\ref{X1X2}),(\ref{X1X2tilde}),(\ref{Xi}), and

376: separating the contributions from the winding modes, one finds

377: \footnote{For a rigorous and detailed derivation of the Hamiltonian see

378: \cite{Russo:1996if}. Here we are interested in outlining only the

379: major steps needed to obtain the mass spectrum.}

380: \bea

381: H_B&=&H_0+H_{int}, \nonumber \\

382: \label{H0}

383: \alpha' H_0 &=& \frac{R_9^2 w_0^2}{2\alpha'}

384: +\frac{1}{2}\sum_{k,m} [P_{(k,m)}^a P_{(-k,-m)}^a + \omega^2_{k\,m}X_{(k,m)}^a X_{(-k,-m)}^a]\\

385: \alpha' H_{int}&=& \frac{1}{4 g^2} \sum_{m,n,p} \sum_{k,l,l'}

386: (pk'-m'k)(nl-ml')

387: X_{(l',n)}^a X_{(k,p)}^a X_{(k',m')}^b X_{(l,m)}^b + \nonumber \\

388: && + \frac{iw_{11}}{g} \sum_{k,m,n} m k^2 X_{2(0,m)}

389: X_{(k,n)}^iX_{(-k,-n-m)}^i,

390: \ea

391: where

392: \bea

393: a,b &=&1, \ldots, 8,11, \;\;\;\; m'=-m-n-p, \; \; \; \; k'=-k-l-l' \nonumber \\

394: g^2 &\equiv & \frac{R_{11}^2}{\alpha'}=4\pi^2 R_{11}^3 T_3, \; \; \; \; \;

395: \omega_{k m} =\sqrt{w_{11}^2 k^2+w_9^2m^2\tau_2^2}, \; \; \;

396: \tau_2=\frac{R_9}{R_{11}}.

397: \ea

398: This is clearly a highly non-linear interacting theory.

399: However, notice that the interacting terms are of order ${\cal O}(\frac{1}{g})$

400: and  ${\cal O}(\frac{1}{g^2})$. In the large $g$ limit, such terms

401: are negligible, and can be dropped.

402: Thus, in the limit $g\rightarrow\infty$, given by

403: \beq

404: \label{limit}

405: R_9,R_{11} \rightarrow \infty, \; \; T_3

406: \rightarrow 0 \; \; \; \; (\text{holding} \;\; \alpha' \;\; \text{and} \;\; \tau_2 \;\;

407: \text{fixed})

408: \eq

409: the Hamiltonian reduces to a system of decoupled

410: harmonic oscillators, and the theory can be solved exactly.

411: In particular, we can quantize the system and determine the mass

412: spectrum.

413: At this point it is important to note that since the states we are

414: interested in considering are BPS, their mass is exact, and can be trusted for all

415: radii. Thus, even though the theory was solved for the special

416: limit (\ref{limit}), such states can be studied more generically,

417: a valuable consequence of the BPS condition.

418: 

419: We should also point out that as long as $w_0=w_9 \, w_{11}\neq0$,

420: the spectrum of the Hamiltonian is discrete. If one was

421: considering $R^{10}\times S^1$, there would be flat directions in

422: the Hamiltonian, causing the membrane to be unstable, and the

423: spectrum would be continuous.

424: 

425: After having dropped $H_{int}$ one can proceed to the quantization

426: of $H_0$. The fields can be expanded in terms of creation and annihilation

427: operators ($a=1,\ldots,8,11 $),

428: \bea

429: X^a_{(k,m)}&=&\frac{i}{\sqrt{2}\,\omega_{(k,m)}} \,

430: [\alpha^a_{(k,m)}+\tilde{\alpha}^a_{(-k,-m)}]\, , \;\;\;\;\;\;

431: P^{\,a}_{(k,m)}=\frac{1}{\sqrt{2}} \,

432: [\alpha^a_{(k,m)}-\tilde{\alpha}^a_{(-k,-m)}]\, , \nonumber \\

433: (X^a_{(k,m)})^\dag &=& X^a_{(-k,-m)}\, , \; \; \; \;(P^{\,a}_{(k,m)})^\dag =P^{\,a}_{(-k,-m)},

434: \;\;\;\;\omega_{(k,m)}=\text{sign}(k)\, \omega_{km}\, .

435: \ea

436: Canonical commutation relations give

437: \beq

438: \left[ X^a_{(k,m)},P^b_{(k',m')}\right] = i

439: \delta_{k+k'}\delta_{m+m'}\delta^{ab},\;\;\;\;

440: \left[ \alpha^a_{(k,m)},\alpha ^b_{(k',m')} \right] = \omega_{(k,m)}\delta_{k+k'}\delta_{m+m'}\delta^{ab}\nonumber  \\

441: \eq

442: and similarly for the $\tilde{\alpha}$'s.

443: The explicit form for the time-dependent part of $X^a$ becomes

444: \beq

445: X^a(\tau,\s,\rho)=x^a+\alpha' p^a \tau + i\sqrt{\frac{\alpha'}{2}}

446: \sum_{(k,m)\neq(0,0)}\frac{e^{iw_{(k,m)}\tau}}{\omega_{(k,m)}}

447: \left[\alpha^a_{(k,m)}e^{ik\s+im\rho}+\tilde{\alpha}^a_{(k,m)}e^{-ik\s-im\rho}

448: \right].

449: \eq

450: The quadratic Hamiltonian (\ref{H0}) takes the form

451: \beq

452: \alpha' H_0= \frac{1}{2}\alpha'(p_9^2+p_{11}^2+p_i^2)+\frac{R_9^2

453: w_0^2}{2\alpha'}+{\cal H},

454: \eq

455: with $i=1,\ldots,8 $ and where ${\cal H}$ contains the contributions from the oscillators,

456: \beq

457: {\cal H} = \frac{1}{2} \sum_{m,k}^\infty \bigl[\alpha^a_{(-k,-m)}\alpha^a_{(k,m)}+

458: \tilde{\alpha}^a_{(-k,-m)}\tilde{\alpha}^a_{(k,m)}\bigr].

459: \eq

460: Letting the momenta in the $X^9,X^{11}$ directions be

461: $p_9=\frac{l_9}{R_9}$ and $p_{11}=\frac{l_{11}}{R_{11}}$, with $l_9,l_{11}$

462: integers,

463: the {\it nine-dimensional} mass operators becomes

464: \beq

465: \label{mass}

466: M^2 =2\, H - p_i^2 =

467: \frac{l_9^2}{R_9^2} + \frac{l_{11}^2}{R_{11}^2} + \frac{R_9^2 w_0^2}{\alpha'^2}

468: + \frac{2}{\alpha'}\,{\cal H}.

469: \eq

470: Clearly, from the eleven-dimensional point of view $\frac{l_9}{R_9}$ and $\frac{l_{11}}{R_{11}}$

471: are just momenta, but they play the role of mass terms in nine

472: dimensions.

473: Schwarz \cite{Schwarz:1995du} showed that the

474: non-oscillating part of the spectrum,

475: $M_0\equiv \frac{l_9^2}{R_9^2} + \frac{l_{11}^2}{R_{11}^2} + \frac{R_9^2 w_0^2}{\alpha'^2}$,

476: matched the corresponding spectrum of the (non-oscillating) type IIB $(p,q)$ string.

477: 

478: As in the case of the string, the level matching conditions

479: for the membrane are obtained from the global contraints

480: \beq

481: P^{(\sigma)}=\frac{1}{2\pi\alpha'}\, \int_0^{2\pi} d\sigma \; \p_\sigma X^a \dot{X}^a

482: \equiv 0, \;\;\;\;\;

483: P^{(\rho)}=\frac{1}{2\pi\alpha'}\, \int_0^{2\pi} d\rho \; \p_\rho X^a

484: \dot{X}^a \equiv 0.

485: \eq

486: These can be re-written in terms of mode operators in the

487: following way,

488: \beq

489: N^+_\s -N^-_\s = w_9 l_9, \;\;\;\;N^+_\rho -N^-_\rho = w_{11} l_{11},

490: \eq

491: where

492: \bea

493: N^+_\s &=& \sum_{m=-\infty}^{\infty}\sum_{k=1}^{\infty}\frac{k}{\omega_{km}}

494: \, \alpha^i_{(-k,-m)}\alpha^i_{(k,m)}, \; \; \;

495: N^-_\s =\sum_{m=-\infty}^{\infty}\sum_{k=1}^{\infty}\frac{k}{\omega_{km}}

496: \, \tilde{\alpha}^i_{(-k,-m)}\tilde{\alpha}^i_{(k,m)}, \nonumber \\

497: N^+_\rho &=& \sum_{m=1}^{\infty}\sum_{k=0}^{\infty}

498: \frac{m}{\omega_{km}}\,

499: \bigl[\alpha^a_{(-k,-m)}\alpha^a_{(k,m)}+

500: \tilde{\alpha}^a_{(-k,m)}\tilde{\alpha}^a_{(k,-m)}\bigr] , \; \; \;

501: \nonumber \\

502: N^-_\rho &=& \sum_{m=1}^{\infty}\sum_{k=0}^{\infty}

503: \frac{m}{\omega_{km}}\,

504: \bigl[\alpha^a_{(-k,m)}\alpha^a_{(k,-m)}+

505: \tilde{\alpha}^a_{(-k,-m)}\tilde{\alpha}^a_{(k,m)}\bigr].

506: \ea

507: 

508: We are interested in considering states that are BPS.

509: If we want to add a wave to the 2-brane background while preserving

510: supersymmetry, we must align it \cite{Russo:1996if} along the momentum

511: direction. Furthermore, for a BPS state one can have only

512: right-moving oscillations.

513: For the special case $l_{11}=0$, these conditions imply

514: that the oscillations are only along the $\sigma$

515: direction, $N_\rho^\pm=0$, and that there are no left-moving oscillators, $N_\s^-=0$.

516: The condition $N_\rho^+=0$ implies that these states are built by

517: applying $\alpha^i_{(-k,0)}$ to the vacuum, which gives $\omega_{km}\rightarrow w_{11} k$.

518: Thus, one finds

519: \beq

520: {\cal H}=w_{11}\,N_\s^+=w_0 l_9,

521: \eq

522: yielding

523: \beq

524: M^2_{BPS}=\frac{l_9^2}{R_9^2}+ \frac{R_9^2 w_0^2}{\alpha'^2}

525: + \frac{2}{\alpha'}\,{w_0 l_9}=\Bigl(\frac{l_9}{R_9}+4\pi^2 T_3\,w_0\,R_9 \, R_{11}\Bigr)^2.

526: \eq

527: For a general state with $l_9,l_{11}\neq 0$, the constraints become \cite{Russo:1996if}

528: \beq

529: \label{constraints}

530: N_\s^+=l_9 w_9, \;\;\;\;\; N_\rho^+=l_{11} w_{11},

531: \eq

532: giving

533: \beq

534: {\cal H}=w_0 \sqrt{l_9^2+l_{11}^2\tau_2^2}.

535: \eq

536: Thus, we obtain the more general mass formula

537: \beq

538: \label{GeneralSpectrum}

539: M^2_{BPS}=\Bigl(\sqrt{\frac{l_9^2}{R_9^2}+\frac{l_{11}^2}{R_{11}^2}}

540: + 4\pi^2 T_3 w_0 R_9 R_{11}\Bigr)^2,

541: \eq

542: matching the spectrum of the oscillating $(p,q)$ string \cite{Russo:1996if}, as

543: anticipated by Schwarz.

544: One can rewrite the mass spectrum in a more convenient form,

545: \beq

546: \label{massR11mod}

547: M^2_{BPS}=m_P^2 \Bigl(\sqrt{\frac{l_9^2}{R_9^2}+\frac{l_{11}^2}{R_{11}^2}}

548: +w_0 R_9 R_{11}\Bigr)^2,

549: \eq

550: where $m_P$ is the Planck mass, and the radii are now dimensionless.

551: We are particularly interested in the case $l_9=0$, yielding

552: \beq

553: \label{massR11}

554: M^2_{BPS}=m_P^2 \Bigl(\frac{l_{11}}{R_{11}}+w_0 \,R_9

555: \,R_{11}\Bigr)^2.

556: \eq

557: We would like to point out that, since the particle numbers $N_\s^+,

558: N_\rho^+$ must be positive, the constraints (\ref{constraints})  imply that $w_9,

559: l_9$ must have the same sign (and similarly for $w_{11},l_{11}$).

560: Thus, the states described by (\ref{GeneralSpectrum}) are always

561: massive \footnote{We would like to thank K. Hori for discussions regarding this point.}

562: for non-trivial quantum numbers.

563: 

564: As we will show in Section \ref{ClTrapping}, such states can

565: play an interesting role on cosmological evolution.

566: However, we are particularly interested in the cosmological consequences of

567: states having a mass of the form

568: \beq

569: \label{ourmass}

570: M^2\sim m_P^2 \Bigl(\frac{|l_{11}|}{R_{11}}-|w_0| R_9 \,R_{11}

571: \Bigr)^2,

572: \eq

573: which can become {\it massless}, and signal an enhancement of symmetry at $R_9\sim R_{11}\sim 1$ (for the

574: case $|l_{11}|=|w_0|)$.

575: 

576: Such states do {\it not} appear in the supermembrane spectrum derived in

577: \cite{Russo:1996if}, however one may expect them to be present in the spectrum of M-theory,

578: possibly in heterotic M-theory.  In fact,

579: heterotic string theory contains

580: states which become massless at enhanced symmetry points, much in the same way as

581: in the bosonic string case.

582: The mass spectrum of the heterotic string with a compact dimension

583: of radius $R$ is of the form

584: \beq

585: \frac{1}{4}M_{\,h}^{\,2}=\Bigl(\frac{p_L^{\,2}}{2}+N_L-1 \Bigr) + \Bigl(\frac{p_R^{\,2}}{2}+N_R-c_R \Bigr)

586: \eq

587: where $p_{L,R}=n/R \pm w \, R$, and

588: for clarity we have separated the contributions from the

589: left-moving and right-moving sides.

590: For the NS sector $c_R=\frac{1}{2}$, while for the R sector

591: $c_R=0$. Thus, the presence of the zero-point energy allows for

592: states of the form

593: \beq

594: \label{radion}

595: M^{\,2} \sim m_s^2\Bigl(\frac{1}{R}-R\Bigr)^2

596: \eq

597: with $R$ the radius of any one of the six compact dimensions, i.e.

598: the {\it radion}.

599: Thus, the heterotic string shows enhancement of symmetry as $R$ approaches the

600: self-dual radius.

601: At the moment we are not aware of any explicit

602: construction of states of the form (\ref{ourmass}), where the

603: pivotal role is played by the {\it eleven-dimensional radius}, i.e. the

604: dilaton. However, in analogy with what happens in the case of the

605: radion, eq. (\ref{radion}), we expect

606: that states of the form (\ref{ourmass})

607: should be found in M-theory on $S^1/\mathbb{Z}_2$.

608: 

609: The lack of such configurations in the derivation of \cite{Russo:1996if}

610: is due to the absence of the zero-point energy in

611: the supermembrane quantization, which is believed to be cancelled

612: by fermionic contributions \cite{Duff:1987cs}.

613: It is also consistent with the fact that such (potentially

614: massless) states are not found in type IIA or IIB string theory, which are

615: obtained from dimensional reduction of 11D M-theory.

616: However, we should mention that supermembrane quantization is still not

617: well understood, and that a full quantum theory of the membrane has

618: proven very difficult to obtain.

619: If, in fact, the zero-point energy were not entirely

620: cancelled by fermions, or if supersymmetry was broken, one

621: could still be able to find states of the form

622: (\ref{ourmass}).

623: Another approach would be to relate

624: known instances of enhanced symmetry to M-theory, through

625: appropriate use of the available dualities, dimensional reduction (and oxidation).

626: Lastly, we mention that considering the supermembrane on backgrounds other than the torus may also offer a resolution,

627: noting for example that Type II strings on K3 (dual to heterotic on $T^6$) also possess enhanced symmetry.

628: 

629: Finding whether enhanced symmetry states of the form (\ref{ourmass}) can be present is an interesting problem in its own

630: right, which we would like to consider in the future.

631: However, for the purpose of this paper, we will simply postulate that

632: these states should exist and restrict ourselves to

633: the study of their cosmological consequences.

634: 

635: \section{Classical Approach to Trapping \label{ClTrapping}}

636: 

637: In the last section we found a set of dyonic solutions

638: whose mass depends explicitly on the dilaton or, from the $11$D

639: perspective, $R_{11}$:

640: \beq

641: \label{heavymass}

642: M^{\,2}=m_p^2\Bigl(\frac{l_{11}}{R_{11}}+w_0 R_9 R_{11}\Bigr)^2.

643: \eq

644: From now on we will set $m_p=1$, and restore explicit $m_p$

645: dependence only when needed.

646: We will now consider the effect of a gas of these dyonic states

647: treated as {\it classical} sources for the vacuum field equations

648: of $11$D SUGRA.  The low-energy effective action for the bosonic

649: degrees of freedom is

650: \beq \label{11daction}

651: S=\frac{1}{2\kappa_{11}^2}\int d^{11}x \sqrt{g}

652: \left[R-\frac{1}{48} G_4^2\right]+\frac{1}{6} \int A_3 \wedge G_4

653: \wedge G_4,

654: \eq

655: where $G_4$ is the anti-symmetric four-form flux

656: $G_4=dA_3$, and $\kappa_{11}^2=8\pi G_{11}$, with $G_{11}$ the

657: eleven-dimensional Newton constant. Anticipating a universe with

658: three large spatial dimensions we consider the following ansatz

659: for the metric and four-form

660: \bea \label{ansatz}

661: ds^2&=&-e^{2A}d\eta^2+e^{2B}d\vec{x}\,^2+e^{2C}d\vec{y}\,^2+e^{2D}dz^2,\\

662: G_4&=& h \, dx^1 \wedge dx^2 \wedge dx^3 \wedge dz,

663: \ea

664: where $\vec{x}=(x_1,x_2,x_3), \, \vec{y}=\{y_m\}, \, m=4,\ldots,9$, and

665: $z$ denotes the coordinate of the eleventh dimension. Notice also that

666: we have introduced exponential scale factors $R_m=e^{\,C}$ and

667: $R_{11}=e^{\,D}$. The functions $A,B,C$ and $D$ depend only on

668: $\eta$, and $h$ is a constant. For this choice of flux both the

669: equation of motion and the Bianchi identity for the gauge field

670: are trivially satisfied.  The remaining equations of motion follow

671: from varying the action with respect to the full metric $g_{MN}$,

672: with $M,N=(0,1,\ldots,9,11)$, \beq

673: R_{MN}-\frac{1}{2}g_{MN}R=\kappa_{11}^2 \left(

674: \frac{1}{12}F_{MOPQ}F_N^{\; \;

675: OPQ}-\frac{1}{96}g_{MN}F^2+T_{MN}^{\; \text{sources}} \right), \eq

676: where the dyon sources (\ref{heavymass})

677: are included through their stress tensor $T_{MN}^{\; \text{sources}}$.

678: 

679: Working in the ideal gas approximation the energy density of these

680: sources can be found from the mass, and is given by

681: \beq

682: \rho=\frac{M}{V}=e^{-3B-6C-D} \left( l_{11} e^{-D}+w_0 e^{C} e^{D}

683: \right).

684: \eq

685: The pressures are then given by

686: \beq \label{pressures}

687: p_{\,3}=-\frac{1}{3}\frac{\partial \rho}{\partial B}-\rho, \;\;\;\;

688: p_{\,6}=-\frac{1}{6}\frac{\partial \rho}{\partial C}-\rho, \;\;\;\;

689: p_{\,11}=-\frac{\partial \rho}{\partial D}-\rho,

690: \eq

691: yielding for the eleven-dimensional pressure

692: \beq

693: p_{\,11}=e^{-3B-6C-D} \left( l_{11} e^{-D}-w_0 e^{C} e^{D}

694: \right).

695: \eq

696: If we take the six dimensions to be at the self-dual radius (i.e. $R_m=1$, or $C=0$)

697: we find that the above pressure will act to stabilize $R_{11}$, corresponding

698: to driving $D \rightarrow 0$.

699: There, the pressure vanishes (for the case $l_{11}=w_0$), and the energy

700: density has a local minimum.

701: We should note that in this analysis the flux would not play a major role, since we are

702: interested in solutions with $B(\eta)$ growing large, and one

703: finds that the flux

704: \footnote{However, the flux will be vital at early times (small $B$) \cite{Friess:2004zk}.}

705: scales as $\sim h^2e^{-6B}$.

706: 

707: Such classical attractors have been studied in many contexts throughout the literature.

708: Examples include gases of wrapped strings and branes

709: (see \cite{Battefeld:2005av,Brandenberger:2005nz,Brandenberger:2005fb, Patil:2005nm}

710: and references within), blackhole

711: attractors \cite{Kaloper:2004yj}, D-brane systems \cite{Abel:2005jx}, and

712: cosmologies involving conifold and flop

713: transitions, e.g., \cite{Mohaupt:2005pa}.  Despite these elegant

714: ideas for stabilizing moduli, a detailed analysis shows that initial conditions

715: play a crucial role.  In the example we have here, one finds that

716: stabilization of $R_{11}$ can be quite generic for fixed $C$.

717: However, as noticed in \cite{Berndsen:2005qq},

718: attempting to stabilize all dimensions simultaneously proves

719: difficult and requires extreme fine-tuning.

720: But perhaps a more serious objection is the {\it validity} of this classical approach.

721: The states (\ref{heavymass}) that we have considered thus far are very heavy for generic values

722: of the moduli.  In fact, even at the stabilization point, the energy of these states

723: is typically near the Planck scale.

724: Thus, the low-energy effective theory that provided us with the equations of motion is no

725: longer the proper

726: framework, given the inclusion of massive states. Furthermore, one

727: cannot arbitrarily add only some, but not all, of the massive

728: states of the strings or branes.

729: 

730: We will turn to a more self-consistent approach in the next section, but it is important

731: to stress that, although the classical approach of this section was naive,

732: it is still interesting to see how the inclusion of truly {\em stringy} states

733: can affect the dynamics.

734: That is, low-energy SUGRA is {\it not} string theory, and the addition of the

735: membrane winding and oscillator contributions can be considered a first correction

736: toward the UV completion of the theory.  Thus, a better

737: understanding of the full string (or membrane) spectrum is a first

738: step towards addressing the string vacuum problem and exploring

739: string phenomenology.

740: However, we do support the viewpoint that this approach is not

741: truly consistent, and whereas in the next section we will find a self-consistent way to

742: include {\em stringy} states in the low-energy theory.

743: 

744: \section{Membrane Production in Time-Dependent Backgrounds \label{QmTrapping}}

745: 

746: In this section we consider the corrections which may result from the dynamics

747: of moduli in the low-energy effective action of string (M-)theory, namely SUGRA.

748: As we have seen, string models contain states whose masses depend explicitly on

749: massless moduli.

750: If the moduli are time-dependent, the masses of such string

751: states change, and can even vanish.

752: Of course, this mechanism is not specific to string theory;

753: the Higgs is a field theory example of a modulus which controls

754: particle masses, however the crucial difference here is the target space of the string is space-time itself.

755: 

756: Typically the states become light as the moduli approach

757: locations where symmetries play a special role (ESPs).

758: Near such points, the states can be quantum produced, and must

759: therefore be incorporated into the effective field

760: theory description.

761: As an example of this {\em stringy} Higgs mechanism, consider the case of a D-brane wrapping a collapsing cycle of the extra dimensions.

762: It was realized some time ago \cite{Strominger:1995cz}, that by including additional light states resulting from the

763: collapse of a wrapped D-brane on a shrinking cycle of the Calabi-Yau, the naive singularities of the conifold geometry could be resolved.

764: The singularity in the moduli space was realized as an artifact of integrating out degrees of freedom that were no longer massive and beyond the cutoff

765: of the effective theory.  We see that dynamics can play a vital role when consider which degrees of freedom to include in the effective theory,

766: since the mass scale can be dynamical given the evolution of background fields (in this case the collapse of the extra dimensions and the brane).

767: 

768: One can identify several corrections to the moduli space approximation

769: resulting from the inclusion of the new light degrees of freedom.

770: Although, in this paper we will concentrate on on-shell production

771: neglecting other types of corrections, which we will mention

772: briefly in the conclusions.

773: As we will see, particle production will be suppressed generically,

774: but it will be non-zero and significant for the enhanced symmetry

775: states mentioned at the end of Section \ref{MthStates}.

776: We will then study the effects of the resulting backreaction on

777: cosmological evolution, and show that it can lead to trapping of

778: the moduli under consideration.

779: Such a trapping mechanism has been previously studied in \cite{Watson:2004aq,Kofman:2004yc}.

780: 

781: \subsection{Time-Dependent Backgrounds}

782: 

783: Let us begin by considering the low energy effective theory for M-theory, i.e. $11$D SUGRA.

784: The effective theory for the massless bosonic fields is given by the action (\ref{11daction}).

785: A class of solutions was found in \cite{Lidsey:1999mc}

786: \bea

787: \label{FullScaleFactors}

788: ds^2&=&-e^{2A}d\eta^2+e^{2B}d\vec{x}\,^2+e^{2C}d\vec{y}\,^2+e^{2D}dz^2, \nonumber \\

789: G_4&=& h \, dx^1 \wedge dx^2 \wedge dx^3 \wedge dz,

790: \eea

791: with the scale factors given by

792: \bea

793: A&=&3B+D \nonumber \\

794: B&=&B_0+\frac{1}{3}\log{\sec{h\eta}}+

795: \frac{q}{2}\log{\frac{1+\sin{h\eta}}{\cos{h\eta}}} \nonumber \\

796: C&=&C_0+\frac{1}{6}\log{\cos{h\eta}} \nonumber \\

797: D&=&D_0-\frac{1}{3}\log{\cos{h\eta}}+\frac{1-3q^2}{6q}

798: \log{\frac{1+\sin{h\eta}}{\cos{h\eta}}},

799: \ea

800: where $B_0,C_0,D_0 \; \text{and} \; |q| \leq 1/\sqrt{3}$ are real constants and $-\frac{\pi}{2h}  \leq \eta  \leq \frac{\pi}{2h}$.

801: Notice that in the limit of no flux, $h=0$, we recover flat space and  $-\infty \leq \eta \leq \infty$.

802: We also note that, although this space-time contains a singularity (located at

803: $-\frac{\pi}{2h}$ for $q>0$),

804: we will only be interested in the evolution away from the singularity

805: \footnote{In \cite{Friess:2004zk}, winding mode creation in this background was studied

806: in order to try to resolve the cosmological singularity.

807: In this paper, however, we will focus on production away from the singularity.}.

808: 

809: We now want to consider the production of membrane states

810: of the type discussed in Section \ref{MthStates},

811: having an effective mass

812: which depends on the radii of the compact dimensions.

813: The first question one may ask is why we expect any production at all, since

814: at low energy scales production of string and membrane states should be strongly

815: suppressed.

816: For the moment let us push forward, keeping this important question in mind.

817: Consider the quantum equation of motion for a string/membrane state labelled by $\chi$.

818: As discussed in \cite{Lawrence:1995ct,Gubser:2003vk} the string constraint equations amount to a

819: wave equation for $\chi$,

820: \be \label{chieom}

821: {\chi}^{\prime \prime}_k+\omega_k^2 \chi_k=0,

822: \ee

823: where a prime denotes the derivative with respect to $\eta$, and $\chi_k$

824: are the Fourier modes of $\chi$.

825: We have removed the friction term by a field redefinition $\chi \rightarrow \chi e^{-3C}$.

826: For simplicity, we will rescale time $\eta \rightarrow \eta h^{-1}$ so that

827: $-\frac{\pi}{2}< \eta < \frac{\pi}{2}$.

828: The time dependent frequency is then given by

829: \be \label{omega}

830: \omega^2=e^{6B+2D}h^{-2}\left(  \frac{k_3^2}{e^{2B}} + \frac{k_6^2}{e^{2C}} + \frac{k_{11}^2}{e^{2D}} +m^2(\eta) +\xi R -e^{-6B-2D}h^2(9C^{\prime \, 2}

831: +3C^{\prime \prime}) \right),

832: \ee

833: where $R$ is the $11$D Ricci scalar, $\xi$ is the coupling of $\chi$ to the

834: space-time background, and $m^2$ represents the contributions to the mass coming

835: from the winding and oscillations of the membrane.

836: Also note that $k_3, k_6$ and $k_{11}$ denote {\it comoving}

837: momentum.

838: It will be more convenient at times to think in terms of the nine-dimensional mass

839: \bea \label{masseff}

840: m^2_{{eff}}&=&\left( k^2_{11}e^{-2D} +m_w^2+m_{\text{osc}}^2 \right) \nonumber \\

841: &=&m_p^2 \left(  l_{11}e^{-D}-\omega_0 e^D e^C \right)^2,

842: \eea

843: since it is when this mass vanishes that we expect significant production.

844: Choosing $B_0=C_0=D_0=0$ for simplicity, these states

845: will become massless (for the case $l_{11}=w_0$) at $\eta=0$, where $D=C=0$.

846: Given the time-dependent frequency (\ref{omega}), with $l_{11}$ and $w_0$

847: left arbitrary, we are now ready to consider the production of the

848: membrane states.

849: 

850: \subsection{The Steepest Decent Method \label{sdm}}

851: 

852: Before considering the case of interest, let us review the

853: standard method for calculating particle production of states with

854: a time-dependent frequency (see e.g. \cite{Birrell:1982ix}). A

855: formal solution to the mode equation (\ref{chieom}) is given by

856: \beq \chi_k = \frac{\alpha_k}{\sqrt{2\omega_k}}e^{-i\int \omega_k

857: d\eta}+\frac{\beta_k}{\sqrt{2\omega_k}}e^{i\int \omega_k d\eta}.

858: \eq The normalization condition for the scalar field is then

859: $|\alpha_k(\eta)|^2-|\beta_k(\eta)|^2=1$, which can be used to

860: write the equation of motion as \bea \alpha_k' &=&

861: \frac{\omega_k'}{2\omega_k} e^{2i\int

862: \omega_k(\eta')d\eta'}\beta_k \nonumber \\ \beta_k' &=&

863: \frac{\omega_k'}{2\omega_k} e^{-2i\int

864: \omega_k(\eta')d\eta'}\alpha_k. \ea Initially we start near the

865: adiabatic vacuum and we have $\beta_k<<1$ and $\alpha_k \sim 1$ so

866: that \beq \label{beta} \beta_k \sim \int d\eta

867: \frac{\omega_k'}{2\omega_k} e^{-2i\int^\eta

868: \omega_k(\eta')d\eta'}, \eq to leading order.  Since $|\beta|^2$

869: gives the number of particles produced, we see that production is

870: suppressed as long as $\omega^\prime \ll \omega$. Conventionally,

871: one defines the dimensionless non-adiabatic parameter

872: ${\omega^\prime}/{\omega^2}$, indicating that particle production

873: becomes significant when

874: \be

875: \frac{\omega^\prime}{\omega^2} \sim 1.

876: \ee

877: In order to calculate the total amount of particles produced we need to estimate the integral in

878: (\ref{beta}).  A method for approximating this integral was found

879: in \cite{Chung:1998bt}, which we will summarize below.

880: 

881: We will be interested in $\omega_k$ of the form $\omega_k=\sqrt{k^2+m^2 f(\eta)}$,

882: where we absorb any time dependence of the background into $f(\eta)$.

883: The poles of the integrand in (\ref{beta}) are also

884: branch points in the complex $\eta$ plane.  Let us

885: assume for simplicity that $\omega_k$ has a single pole at

886: $\eta^\star$.

887: Then, near the branch point, the integral in the exponent of (\ref{beta})

888: can be expanded in the following way:

889: \beq

890: \int_{\eta_{in}}^\eta \omega_k(\eta') d\eta' =\int_{\eta_{in}}^\eta \omega_k(\eta')

891: d\eta' + \frac{2M}{3}\sqrt{f'(\eta^\star)}(\eta-\eta^\star)^{3/2},

892: \eq

893: keeping the leading term in the expansion of $f(\eta)$ about $\eta^\star$.

894: We then find

895: \beq

896: \beta_k = \Bigl( \frac{1}{4}\int_{f^\star}\frac{d\delta}{\delta}e^{\frac{-4iM}{3}

897: \sqrt{f'(\eta^\star)}\delta^{3/2}} \Bigr) \,

898: e^{-2i\int_{\eta_{in}}^{\eta^\star} \omega_k(\eta')d\eta'},

899: \eq

900: where the $f^\star$ denotes the steepest descent contour.

901: The factor in front of the exponential can be shown to equal

902: $\frac{i\pi}{3}$, giving us

903: \beq

904: \beta_k = \frac{i\pi}{3} e^{-2i\int_{\eta_{in}}^{r} \omega_k(\eta')d\eta'} \,

905: e^{-2i\int_{r}^{\eta^\star} \omega_k(\eta')d\eta'},

906: \eq

907: where $\eta^\star=r-i\mu$.

908: The first integral in the expression

909: above is real, making the first exponential a pure phase.

910: The second integral can be approximated by

911: $\int_{r}^{\eta^\star} \omega_k(\eta')d\eta' \sim i \pi \mu

912: \omega_k(r)$, yielding

913: \beq \label{betaform}

914: |\beta_k|^2 = \biggl(\frac{\pi}{3}\biggr)^2 e^{-\pi\mu\omega_k(r)}.

915: \eq

916: 

917: Thus, to calculate the leading contribution to particle production

918: it is sufficient to identify the real and imaginary contributions

919: to the zeros of $\omega_k$. Because of the choice of integration

920: contour, we should note that we are only interested in the zeros which are in the

921: lower half plane.

922: 

923: \subsection{Membrane Production}

924: 

925: We now want to determine which conditions will lead to significant production

926: in the chosen background (\ref{FullScaleFactors}).  The

927: adiabatic approximation is valid as long as ${\omega^\prime}/{\omega^{2}}<1$,

928: and as long as this condition holds production will be insignificant.

929: Thus, the first step is to examine when the non-adiabaticity is appreciable for our frequency

930: \be \label{ourfreq}

931: \omega^2_k=e^{6B+2D}h^{-2}\left(  \frac{k_3^2}{e^{2B}}  +m_{eff}^2(\eta)

932: +\xi R -e^{-6B-2D}h^2(9C^{\prime \, 2}+3C^{\prime \prime}) \right),

933: \ee

934: where we will neglect any momentum in the extra dimensions (i.e. we take $k_6=0$),

935: but keep $k_{11}$ non-zero.

936: 

937: One obvious location where non-adiabaticity becomes large is near the cosmological

938: singularity ($\eta=-\pi/2$), which was the case examined in \cite{Friess:2004zk}.

939: Here, however, we are not interested in the region near the

940: singularity. Thus, as long as we are not too close to the singular

941: region, we can still construct the appropriate in-vacuum

942: \footnote{This is analogous to the calculation of production at the end of inflation,

943: where the initial big-bang singularity is irrelevant to the calculation.

944: See \cite{Birrell:1982ix} for a detailed discussion on the method of constructing

945: an asymptotic, adiabatic vacuum in such singular space-times.}.

946: Let us now examine the behavior of the non-adiabatic parameter away from

947: the initial singularity.  We expect that away from the values $C=D=0$ the mass term

948: should dominate over all other terms in the frequency (\ref{ourfreq}).

949: In this limit, $$\frac{\omega^\prime}{\omega^2} \approx \frac{m_{eff}^\prime}{m_{eff}^2},$$

950: and we expect the non-adiabaticity to be peaked near $m_{eff}\approx 0$, which is where

951: the states become massless.

952: 

953: In Figures \ref{fig1}, \ref{fig2}, and \ref{fig3}, we present the behavior of the

954: {\it exact} non-adiabatic parameter for various values of the background and momenta.

955: In all cases we find that the production is sharply peaked around $\eta \approx 0$.

956: As noted in \cite{Kofman:2004yc} (and references within), this means that production can

957: be treated as an instantaneous event, making the calculation of production

958: and its backreaction much more tractable.  We see from the figures that

959: away from $\eta \approx 0$ the adiabatic vacuum is an excellent approximation,

960: and that well-defined in and out regions do indeed exist.

961: 

962: 

963: \begin{figure}

964: \includegraphics[width=10.5cm]{fig1}

965: \caption{\label{fig1} The non-abiabatic parameter ${\omega^\prime}/{\omega^{2}}$

966: for different values of $\xi$, with

967: $\xi=0$ (top curve), conformal coupling $\xi=9/40$ (middle curve), and $\xi=5$ (bottom curve).

968: As $\xi$ increases, non-adiabaticity is suppressed.

969: Note that the non-adiabticity is

970: concentrated near $\eta \approx 0$ for all choices of $\xi$.}

971: \end{figure}

972: %

973: \begin{figure}[!]

974: \includegraphics[width=9.5cm]{fig2}

975: \caption{\label{fig2} The non-adiabatic parameter ${\omega^\prime}/{\omega^{2}}$ for different

976: values of the flux $h=1/150$ (top curve), $h=1/100$ (middle curve), and $h=1/50$ (bottom curve).}

977: \end{figure}

978: %

979: %

980: \begin{figure}[!]

981: \includegraphics[width=9.5cm]{fig3}

982: \caption{\label{fig3} The non-adiabatic parameter ${\omega^\prime}/{\omega^{2}}$ for different values of the

983: momentum $k_3=1/500$ (top line), $k_3=1/100$ (middle line), and $k_3=1/50$ (bottom line).}

984: \end{figure}

985: %

986: 

987: One point of concern is the effect of the time-dependent geometry on the

988: effective mass  $$m^2_{total}=m_{eff}^2+\xi R + \text{gravity terms}$$ through the

989: last three terms in (\ref{ourfreq}).

990: As seen in Fig. \ref{fig1}, the terms including the coupling to the geometry $\xi$,

991: have the effect of suppressing the amount of production, but this suppression is

992: negligible for reasonable values of the coupling.  In particular, the cases of minimal

993: ($\xi=0$) and conformal $(\xi=9/40)$ coupling have little or no effect.  This can be understood

994: by examining the Ricci curvature which, for small flux, is negligible away from the

995: singular region (i.e., $R \sim h^2$).  The requirement of small flux can be seen

996: from Fig. 2, where we find that production will be suppressed unless $h<m_p$.

997: Fig. 3 shows a similar result for the momenta, $k_3 < m_p$.

998: Let us now consider the production from a more quantitative perspective.

999: 

1000: We have seen that the non-adiabaticity is focused near $\eta \approx 0$.  Using this

1001: and the effective mass (\ref{masseff}), we can expand the frequency (\ref{ourfreq})

1002: for small $\eta$

1003: \be \label{approxomega}

1004: \omega_k^2(\eta)=f_0+f_1 \eta+f_2 \eta^2,

1005: \ee

1006: where, after restoring the explicit dependence on $m_p$ ,

1007: \bea

1008: f_0&=& \frac{k_3^2+a_0m_p^2+\frac{1}{2} h^2}{h^2}, \;\;\;\;\;\;\;

1009: f_1 = \frac{a_1 k_3^2+a_2 m_p^2}{h^2}, \;\;\;\;\;\;\;

1010: f_2 = \frac{a_3k_3^2+a_4m_p^2+\frac{1}{4} h^2}{h^2}, \nonumber \\

1011: a_0&=&(w_0-l_{11})^2,

1012: \;\;\;\;a_1=\frac{1+3q^2}{3q}, \;\;\;\;

1013: a_2=\frac{2w_0(w_0-l_{11})+3q^2(w_0^2-4w_0l_{11}+3l_{11}^2)}{3q} \nonumber \\

1014: a_3&=&\frac{1+24q^2+9q^4}{18q^2}, \;\;

1015: a_4=\frac{2w(w-l)+3q^2(13w^2-23wl+6l^2)+q^4(9w^2-72wl+91l^2)

1016: }{18 q^2}.\nonumber

1017: \ea

1018: Following the method outlined in Section \ref{sdm}, we proceed by finding the zeros of (\ref{approxomega}).

1019: The frequency vanishes at

1020: \beq

1021: \eta^*=\frac{-f_1\pm\sqrt{f_1^{\, 2}-4\, f_0 \, f_2}}{2f_2}.

1022: \eq

1023: Remembering that we are interested in the zeros in the lower

1024: half plane, we choose

1025: \beq

1026: \eta^\star=-\frac{f_1}{2f_2}-i\frac{\sqrt{4f_0

1027: f_2-f_1^2}}{2f_2}\equiv r-i\mu,

1028: \eq

1029: where we have $4f_0 f_2-f_1^2>0$.

1030: The amount of particle production for a given mode is then given by (\ref{betaform}),

1031: and is found to be

1032: \beq

1033: |\beta_k|^2 = \biggr(\frac{\pi}{3}\biggr)^2 e^{-\pi\frac{4f_0

1034: f_2-f_1^2}{4 \, f_2^{3/2}}}.

1035: \eq

1036: Moreover, using the fact that $m_p\gg k_3,h$, we find

1037: \beq \label{betaeqn}

1038: |\beta_k|^2 = \biggr(\frac{\pi}{3}\biggr)^2 e^{-\frac{A_1\,m_p}{h}}

1039: e^{-\frac{A_2\,k_3^2+A_3\,h^2}{m_p h}+{\cal

1040: O}(\frac{1}{m_p^2})},

1041: \eq

1042: where $A_1, A_2, A_3$ are constants that depend on $w_0, l_{11},

1043: q$.

1044: 

1045: Given this expression we can now return to the initial concern that string and

1046: membrane particle production should suffer Planck scale suppression

1047: at low energies.  This can be seen by the suppression factor $e^{-\frac{A_1\,m_p}{h}}$

1048: appearing in (\ref{betaeqn}), which is related to the usual suppression

1049: due to the Hagedorn density of states (see e.g. \cite{Gubser:2003vk}).

1050: However, in this case we find that $A_1 \propto (l_{11}-w_0)^3$, which vanishes

1051: when $l_{11}=w_0$, consistent with the condition for having

1052: massless states\footnote{This is the motivation for wanting states

1053: having a mass of the form (\ref{ourmass}).}

1054:  at the self-dual radius $D=0$.  Thus, the enhanced symmetry results in

1055: additional light states which will be copiously produced.

1056: In fact, we find significant particle production when $l_{11}=w_0$,

1057: given by

1058: \beq \label{betaeqna}

1059: |\beta_k|^2 = \biggr(\frac{\pi}{3}\biggr)^2 e^{-\frac{3\pi}{2\,m_p\,h}

1060: \bigl(\frac{h^2+2 k_3^2}{l_{11}\gamma}

1061: \bigr)},

1062: \eq

1063: where $\gamma\equiv \frac{1-3q^2}{q}$.

1064: Furthermore, we will choose $l_{11}=w_0=1$, since states

1065: with higher winding and momentum numbers would decay to this configuration.

1066: We can immediately see from (\ref{betaeqna}) the same behavior that we found for the non-adiabatic parameter in the previous section.

1067: Namely, we see that for $h, k_3 < m_p$ production will result and above the Planck scale production will cease.  This behavior agrees with the numerical studies

1068: of the non-adiabatic parameter, which can be seen in Fig. \ref{fig2} and Fig. \ref{fig3}.

1069: 

1070: From (\ref{betaeqna}) we can calculate the eleven-dimensional number

1071: density at the creation time $\eta^*$ by summing over all modes

1072: \beq \label{densities}

1073: n_\chi (\eta^*)

1074: =\frac{1}{V_0^{(7)}} \int_0^\infty \frac{d^3k_3}{(2\pi)^3}

1075: \; |\beta_k|^2,

1076: \eq

1077: $V_0^{(7)}$ is the dimensionful volume of the extra dimensions

1078: at the time the particles were produced \footnote{We

1079: note that in the calculation of the 11D densities there is an

1080: implicit integration over the momenta of the extra dimensions (and the Kaluza-Klein modes).

1081: However, we have chosen to work with the $l_{11}=1$ states

1082: and any additional momentum (e.g. $k_6$) is taken to vanish.}.

1083: The energy density of created particles is given by

1084: \beq \label{edense}

1085: \rho= \frac{1}{V_0^{(7)}} \int_0^\infty \frac{d^3k_3}{(2\pi)^3} \;  \omega_k

1086: |\beta_k|^2.

1087: \eq

1088: Finally, the pressure can be found from the energy density as in (\ref{pressures}),

1089: and the effective stress energy tensor of the produced membrane states is then

1090: \be \label{st}

1091: T^{M}_{\; \; N}=diag \left( -\rho, p_3, p_6, p_{11} \right).

1092: \ee

1093: 

1094: \section{Cosmology, Backreaction, and Moduli trapping \label{cosmo}}

1095: 

1096: We now want to consider the effect of the states produced in the last section

1097: on the evolution

1098: of the background (\ref{FullScaleFactors}).

1099: One possible way to include the effects of such states would be

1100: to study the evolution of

1101: linear perturbations about the background (\ref{FullScaleFactors}).

1102: However, one finds that this approach is not adequate, since the presence of the states

1103: alters the evolution drastically, making the effect of order one.

1104: Instead, we consider the new metric

1105: \be \label{newmetric}

1106: ds^2=-dt^2+a(t)^2 dx_3^2 + b(t)^2 dx_6^2+R_{11}^2(t) dx_{11}^2,

1107: \ee

1108: where we will work in synchronous gauge, and we have assumed that the perturbed metric will

1109: respect the symmetries (topology) of the original metric\footnote{

1110: Note that for simplicity we will take six of the radii to be equal (i.e., $R_m=b$

1111: with $m=4,5,6,7,8,9$).

1112: This could easily be generalized without changing our conclusions.}.

1113: Introducing the Hubble parameters $H_3=\frac{\dot{a}}{a}, H_6=\frac{\dot{b}}{b}$ and

1114: $H_{11}=\frac{\dot{R}_{11}}{R_{11}}$, the equations of motion can be written as

1115: \be \label{theeom}

1116: \dot{H}_i=H_i \sum_{j} H_j + 8\pi G_{11} \left( p_i +\frac{1}{10}

1117: \left(  \rho- \sum_{j} p_j \right) \right),

1118: \ee

1119: with $i,j=1 \ldots 9,11$ running over all spatial dimensions.  We have again taken

1120: the matter sources

1121: to be in the form of a perfect fluid characterized by their energy density and pressures,

1122: and having a stress tensor given by

1123: \be \label{sth}

1124: T_{MN}=diag \left( \rho, a^2 p_3, b^2 p_6, R_{11}^2 p_{11} \right).

1125: \ee

1126: 

1127: The constraint equation

1128: \be

1129: \sum_{i<j} H_i H_j =8\pi G_{11} \rho

1130: \ee

1131: and the equations motion allow one to obtain the continuity equation

1132: \be

1133: \dot{\rho}=-\sum_i \left( \rho + p_i \right) H_i.

1134: \ee

1135: In general these equations are very difficult to solve.

1136: However, in the absence of matter (i.e., $\rho=p_i=0$) one finds the well-known

1137: Kasner solutions,

1138: \bea

1139: a(t)\sim t^{c_1}, \;\;\;\;\;\;  b(t)\sim t^{c_2}, \;\;\;\;\;\;\;  R_{11}(t)\sim t^{c_3},

1140: \nonumber \\

1141: \text{with} \;\;\;\sum_i c_i = \sum_i (c_i)^2=1.

1142: \eea

1143: Thus, in a universe initially filled with radiation or matter the expansion will

1144: dilute the sources,

1145: and we expect the late-time behavior to approach that of the Kasner solution.

1146: 

1147: We want to include the effect of the produced states on the background

1148: (\ref{FullScaleFactors}).

1149: As we found in the last section we can include the states through their energy

1150: density (\ref{edense}).

1151: It is important to note that number density is only well defined in the adiabatic

1152: out region.

1153: That is, near the region of non-adiabaticity, particle number (and thus energy density)

1154: can fluctuate.

1155: However, following \cite{Kofman:2004yc} we will assume that after

1156: the first pass through the production point no further production occurs, and the particle

1157: number remains fixed.

1158: This approximation is certainly acceptable, since the exact treatment would merely increase

1159: the number of particles, enhancing the trapping mechanism we will discuss.

1160: 

1161: Perhaps a more serious drawback is the validity of the SUGRA approximation near the

1162: production point.

1163: For the background we have considered, production will occur near the Planck scale at the

1164: center of the M-theory moduli space.

1165: From the ten-dimensional point of view, this will be a fixed point of S-duality where a

1166: perturbative description of the theory is not available at this time.

1167: However, we can trust our equations away from the non-adiabatic region, and we are primarily

1168: interested in the attractive behavior towards this region.

1169: Thus, we will consider the effects coming from production, remembering

1170: that further corrections may become important near the non-adiabatic region.

1171: In the worst case, we will see that the corrections we consider lead us naturally to the

1172: non-adiabatic region, where a further knowledge of M-theory dynamics is needed.

1173: 

1174: Using the expression (\ref{betaeqna}) in (\ref{densities}) we find for the number density

1175: at the time of creation $\eta^{*}$

1176: \bea

1177: n_\chi(\eta^{*}) &=& \frac{1}{\vol}

1178: \int_0^\infty \frac{d^3k}{(2\pi)^3} \; \left( \frac{\pi}{3} \right)^2

1179: e^{ -\frac{3\pi}{h\gamma m_p} \left( k_3^2+ \frac{1}{2}h^2 \right) } \\

1180: &=& \frac{\sqrt{3}\gamma^{3/2}}{648 \, \pi}  \left(m_p \, h\right)^{5}

1181: e^{-\frac{3\pi h}{2 \gamma m_p}},

1182: \eea

1183: where we used $V_0^{(7)}=(h\,m_p)^{-7/2}$.

1184: This is the number density of particles resulting from production

1185: at the time when the scale factors pass near the region of non-adiabaticity $D=C=0$.

1186: %

1187: In terms of the new metric (\ref{newmetric}) the creation time

1188: will be denoted by $t_0$, and the production point corresponds to $R_{11}=b=1$.

1189: Away from the non-adiabatic region the number of particles will remain

1190: constant, and the density in the new metric is given by

1191: \beq

1192: n_\chi(t)=\left( \frac{a^3(t_0) \, b^6(t_0) \, R_{11}(t_0)}{a^3(t)\, b^6(t)\, R_{11}(t)}

1193: \right)n_\chi(t_0)\equiv \frac{V(t_0)}{V(t)}n_\chi(t_0).

1194: \eq

1195: 

1196: At this point it is convenient to return to setting

1197: $m_p=1$; we will reinstate the explicit dependence on the Planck

1198: mass when necessary.

1199: The energy density at time $t_0$ is given by

1200: \beq

1201: \rho(t_0) =\frac{\pi^2}{9 \vol} \int_0^\infty \frac{d^3k}{(2\pi)^3} \;

1202: \omega_k \, e^{ -\frac{3\pi}{h \gamma} \left( k_3^2+ \frac{1}{2}h^2 \right) }.

1203: \eq

1204: The frequency $\omega_k$ evaluated in the new background takes the

1205: form

1206: \be

1207: \omega_k^2 \sim \frac{k_3^2}{a^2}+ \left(  \frac{1}{R_{11}}-R_{11} b \right)^2

1208: \equiv \frac{k_3^2}{a^2}+ m_{eff}^2,

1209: \ee

1210: where we have dropped gravitational terms which have been shown to be

1211: of higher adiabatic order and are negligible (recall that these terms scale like $h^2$, and

1212: $h^2 \ll m_{eff}^2$).

1213: Using this frequency we find that the energy density at $t_0$ is

1214: given by

1215: \beq

1216: \rho(t_0)= \frac{\pi^2}{9\vol}  \int_0^\infty \frac{d^3k}{(2\pi)^3}

1217: \sqrt{\frac{k_3^2}{a^2}+ m_{eff}^2} \;

1218: e^{ -\frac{3\pi}{h \gamma} \left( k_3^2+ \frac{1}{2}h^2 \right)}.

1219: \eq

1220: At a later time $t$, the energy density scales in the same way as

1221: the number density,

1222: \beq

1223: \label{rhogo}

1224: \rho(t)=\frac{V(t_0)}{V(t)} \frac{a \, \gamma \,h^{9/2} }

1225: {144 \pi}  \;  m_{eff}^2  \; e^{-\frac{3\pi}{2\gamma}\left( h-a^2 h^{-1}

1226: m_{eff}^2\right)} \mathcal{K}_1\left(  a^2 h^{-1}m_{eff}^2 \right),

1227: \eq

1228: where $\mathcal{K}_1$ is a Bessel function of the second kind.

1229: 

1230: \subsection{Cosmological Evolution and Trapping}

1231: 

1232: We are interested in the dynamics after passing through the ESP at $R_{11}=b=1$.

1233: Notice that in the absence of matter the background solution (\ref{FullScaleFactors})

1234: predicts that the radii (moduli) will continue to evolve to larger values.

1235: We will demonstrate that, by including the states produced at the ESP via their stress energy

1236: tensor (\ref{st}),

1237: the motion of the radii will be reversed back towards the ESP, and the moduli can eventually

1238: become trapped.

1239: We again note that there should also be a contribution to the stress tensor coming from the flux,

1240: but such terms scale as $\rho \sim p_i \sim \frac{h^2}{a^6}$ and are therefore only important at

1241: early times (i.e. they are red-shifted by the expansion of $a(t)$).

1242: 

1243: The evolution begins in the adiabatic region, where $R_{11}>1$, $b>1$ and the mass term

1244: dominates the energy.

1245: In this limit the energy density (\ref{rhogo}) becomes

1246: \be \label{rhoapprox}

1247: \rho \approx  m_{eff} \, n_\chi(t).

1248: \ee

1249: In the opposite limit, near $R,b \approx 1$, the mass is negligible and the energy density becomes

1250: \be \label{rhoapprox2}

1251: \rho \approx \frac{n_\chi(t)}{a}.

1252: \ee

1253: From (\ref{rhoapprox}) we can determine the pressure in the respective dimensions using (\ref{pressures}).

1254: We find that for $R_{11}>1$ and $b>1$ these scale as

1255: \bea

1256: p_{\,3}&\approx&0, \;\;\;\;\;

1257: p_{\,6} \approx-\frac{n_\chi(t)}{6} R_{11}b, \;\;\;\;\;\;

1258: p_{\,11}\approx-n_\chi(t)R_{11}b \;\;\;\;\;\;\;\;

1259: \left( R_{11}>1 \; \text{and} \; b > 1\right).

1260: \eea

1261: We see that for large $R_{11}$ (or large $b$), the negative pressure due to the wrapped M2 branes

1262: dominates.

1263: From the equations of motion (\ref{theeom})

1264: we see that as $R_{11}$ grows large the negative source terms will dominate over

1265: the first terms

1266: on the right side of (\ref{theeom}), given that $H_{11}$ is not too large.

1267: At the turning point $H_{11}=0$ and we find that $\dot{H}_{11}<0$, due to the large negative pressure $p_{11}$,

1268: showing that $R_{11}$ reaches a

1269: maximum value and

1270: turns back toward the ESP \footnote{It is important to note that in anisotropic space-times

1271: negative pressures lead to contraction, not accelerated expansion.}.

1272: After the radii pass through the ESP for a second time, we have $R_{11}<1$, and the

1273: pressures are then given by

1274: \bea \label{pb}

1275: p_{\,3}&\approx&0, \;\;\;\;\;\;\;

1276: p_{\,6}\approx \frac{n_\chi(t)}{6} R_{11}b, \;\;\;\;\;\;\;

1277: p_{\,11}\approx  \frac{n_\chi(t)}{R_{11}} \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;  \left(R_{11}<1\right),

1278: \eea

1279: In the last step we have assumed that the radii have moved sufficiently below the ESP so that we are again

1280: in an adiabatic region, and (\ref{rhoapprox}) still

1281: gives the relevant energy density\footnote{Strictly speaking this is not quite correct.

1282: As the radii pass back through the ESP (non-adiabatic region), further particle production

1283: is possible and the density of particles can increase.  However, including the production of

1284: additional states will only act to enhance the trapping mechanism that we will discuss.

1285: See \cite{Kofman:2004yc} for a related discussion.}.

1286: We see from (\ref{pb}) that as $R_{11}$ continues to evolve towards smaller values, the

1287: pressure becomes large and positive. From

1288: (\ref{theeom}) this means that $R_{11}$ will reach a minimum value and

1289: go back towards the ESP.

1290: This behavior will continue and $R_{11}$ will oscillate around the ESP, with the oscillations

1291: damping due to the expansion of $a(t)$.

1292: 

1293: However, one place we have been cavalier is in the evolution of the extra dimensional scale factor $b$.

1294: In fact, from its associated

1295: pressure (\ref{pb}) we see that for $b \ll 1$ the pressure will vanish.

1296: Thus, since there is no pressure to

1297: prevent its collapse, $b$ will continue to run to smaller values.

1298: Furthermore, we have over-simplified the entire evolution by assuming that $R_{11}=1$ and $b=1$

1299: occur simultaneously.  If this were the case it would be the result of extreme fine-tuning.

1300: \begin{figure}

1301: \includegraphics[width=10.5cm]{fig4}

1302: \caption{\label{fig4} Evolution of $R_{11}$ and $b$ for the exact density (\ref{rhogo}) and pressures, including only the membrane sources related to $R_{11}$.}

1303: \end{figure}

1304: Instead, as can be seen in Fig. \ref{fig4}, we find that for differing values of $R_{11}$ and

1305: $b$ trapping of $R_{11}$ can occur away from the ESP at a value that is determined by the

1306: running of $b$ to its asymptotic value\footnote{We note that upon compactification to

1307: ten dimensions this is the result found in \cite{Watson:2003gf}, where it was shown that the

1308: radii could be stabilized at the expense of a running dilaton}.

1309: 

1310: Despite the evolution of $b$, we can remedy the situation by simply including the other membrane

1311: states discussed in Section \ref{MthStates}.  That is, near $b=1/\sqrt{R_{11}}$ there are additional

1312: massless states that can be produced with masses

1313: \be

1314: m^{(b)}_{eff}= \left|  \frac{1}{b}-R_{11}(t) b(t) \right|,

1315: \ee

1316: where now the momentum is taken in the $b$ direction (i.e., $k_9 \neq 0$, $k_{11}=0$).

1317: The production of these states is handled analogously

1318: to the previous states, and leads to an additional contribution to the energy density

1319: \be

1320: \rho^{(b)}=6 \, \tilde{n}_\chi(t) \, m_{eff},

1321: \ee

1322: where the factor of six comes from considering states produced equally in all six dimensions.

1323: These states provide the needed pressure term

1324: at small $b$

1325: \bea \label{pressb}

1326: p^{(b)}_6&=& \frac{\tilde{n}_\chi(t)}{b} \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;

1327: \left( b<1  \right),

1328: \eea

1329: which will cause the motion of $b$ to return to the ESP.

1330: 

1331: Given the additional sources, we now expect from the asymptotic behavior

1332: that the moduli should be trapped.

1333: However, the dynamics is actually quite involved due to the presence of non-linearities.

1334: Using the experimental result that $H_3>0$ (i.e. we live in three large dimensions),

1335: we examine the system numerically with the results appearing in Figures \ref{fig5} $-$ \ref{fig7}.

1336: In Fig. \ref{fig5}, we see the evolution of the radii  $R_{11}$ and $b$ given initial conditions

1337: consistent with $H_3>0$.

1338: \begin{figure}

1339: \includegraphics[width=10.5cm]{fig5}

1340: \caption{\label{fig5} Evolution of $R_{11}$ (dark curve) and $b$ (light curve) for the exact density (\ref{rhogo}) and pressures, including membranes in all extra dimensions.}

1341: \end{figure}

1342: %

1343: \begin{figure}

1344: \includegraphics[width=10.5cm]{fig6}

1345: \caption{\label{fig6} $R_{11}$ for differing initial values of the expansion rate, with $\dot{R}_0=(0.001, 0.5)$ being barely distinguishable and with $\dot{R}_0= 0.6$ (light curve) we see the period begins to grow.  We see that as the initial expansion rate is increased trapping takes longer to occur with $H_{11}>m_p$ leading to the case of no trapping.}

1346: \end{figure}

1347: %

1348: \begin{figure}

1349: \includegraphics[width=10.5cm]{fig7}

1350: \caption{\label{fig7} Evolution of $a(t)$ compared to $t^{1/2}$.}

1351: \end{figure}

1352: The jagged oscillations along the curve of $R_{11}$ are not due to numerical error,

1353: but rather

1354: result from the coupling to $b$ and the discontinuities associated with the pressure

1355: changing sign.

1356: We find that the radii will continue to oscillate with a decreasing amplitude, due to

1357: the expansion of $a(t)$.

1358: At late times we find that the radii $R_{11}$ and $b$ approach a constant value, which from (\ref{rhoapprox2}) implies

1359: $\rho \sim a^{-4}$ and $a(t) \sim t^{1/2}$.

1360: That is, our three dimensional universe evolves to that of a radiation dominated universe and

1361: the radii are trapped near the ESP $R_{11}=b=1$.

1362: We find that the trapping is robust, given that the initial expansion rates of the radii do

1363: not exceed the Planck scale (i.e., $H_6<m_p$ and $H_{11}<m_p$.  This condition can be seen

1364: from Fig. \ref{fig6}, where we have plotted the evolution of $R_{11}$ for increasing values of

1365: the initial expansion rate $H_{11}$, and a similar result follows for $b(t)$.

1366: In Fig. \ref{fig7}

1367: we have presented a comparison of the late time behavior of $a(t)$ versus that of a radiation

1368: dominated universe $a(t)\sim t^{1/2}$.  The wiggles in the evolution of $a(t)$ before the radii

1369: completely stabilize, naively may suggest the possibility of cosmological signatures coming from

1370: the trapping mechanism.

1371: 

1372: \section{Conclusions and Future Prospects}

1373: 

1374: We have considered both classical and quantum corrections to

1375: low-energy M-theory coming from dynamics of BPS membrane bound

1376: states whose effective mass depends on the radii of the extra

1377: dimensions. Including such states classically leads to an

1378: attractor mechanism that fixes moduli but is inconsistent with the

1379: use of the effective field theory approach.

1380: 

1381: Insisting on an effective field theory description, we then

1382: consider quantum mechanical production of these membranes in a

1383: time-dependent background, and find the expected result that the

1384: production suffers Planckian suppression. However, we do find the

1385: possibility of significant and finite production for states that

1386: exhibit enhanced symmetry. We believe that these should correspond

1387: to non-threshold bound states of membranes and gravitational

1388: waves, which become tensionless at the eleven-dimensional

1389: self-dual point, $R_{11}=l_p$. An exact construction of such

1390: states is challenging, due to the problems of quantizing the

1391: supermembrane and the lack of an effective description of the

1392: theory in this regime.

1393: 

1394: In \cite{Russo:1996if} it was shown that one can obtain the BPS

1395: spectrum of the supermembrane in special limits and by utilizing

1396: properties of BPS configurations. However, the enhanced states we

1397: are interested in require non-vanishing vacuum energy, which is

1398: not compatible with the $\mathcal{N}=2$ SUSY case considered in

1399: \cite{Russo:1996if}. Therefore, guided by the analysis of

1400: \cite{Russo:1996if}, we conjecture the existence of the enhanced

1401: states, awaiting a more concrete construction. One possibility for

1402: their existence is the case of $\mathcal{N}=1$ heterotic M-theory,

1403: in analogy with the enhanced gauge symmetry of heterotic strings.

1404: Another intriguing possibility is the recent conjecture that the

1405: spectrum of string / M-theory contains states that lie below the

1406: BPS bound \cite{Arkani-Hamed:2006dz} .

1407: 

1408: Assuming that such enhanced states exist, we find that they can

1409: have a critical impact on the evolution of moduli, which would

1410: have been missed in the naive low-energy effective theory

1411: neglecting dynamics. We have found that, by including the

1412: backreaction of the membrane states produced, the radii are

1413: dynamically attracted to values near fixed points of S- and T-

1414: duality. This effect would be missed if one did not have a

1415: knowledge of (enhanced) M-theory states having masses which depend

1416: on evolving moduli. Thus, the lesson we learn is that the presence

1417: of time-dependence introduces a dynamical mass scale that must be

1418: taken into careful consideration in the effective field theory.

1419: Furthermore, it is paramount not to forget the string theory

1420: origin of the low-energy effective action.

1421: 

1422: In addition to the corrections from on-shell particle production

1423: which we have explored here, there will be radiative corrections

1424: coming from the presence of the light membrane states. In fact, it

1425: was shown in \cite{Silverstein:2003hf} that, for the case of

1426: colliding D-branes, open strings becoming light lead to

1427: corrections to the gauge theory propagator resulting in a speed

1428: limit for the moduli. In the case we have considered here a

1429: similar story should hold, but this time it is the closed string

1430: moduli which should be slowing down (the radii of the extra

1431: dimensions). This offers a challenge, since the gauge theory

1432: interpretation of such processes is unclear. We find the

1433: possibility of speed limits for radii intriguing, and we hope to

1434: report on it shortly.

1435: 

1436: 

1437: \begin{acknowledgments}

1438: We would like to thank R. Brandenberger, B. Holdom, L. Kofman, A. Krause, D. Lowe, L. McAllister,

1439: S. Patil, D. Podolsky, S. Prokushkin, and H. Nastase for useful discussions.  We would especially like to

1440: thank A. Jevicki, J. Russo, and A. Tseytlin for critical comments and suggestions and we are

1441: grateful to K. Hori for finding an error in an earlier draft.  S.C. would like to thank the

1442: University of Toronto for hospitality.  This work was financially supported in part by the

1443: National Science and Engineering Research Council of Canada, the Schnabel Woods institute,

1444: and the U.S. Department of Energy under contract DE-FG02-91ER40688, TASK A.

1445: \end{acknowledgments}

1446: \newpage

1447: \begin{thebibliography}{99}

1448: 

1449: \bibitem{Vafa:2005ui}

1450:  C.~Vafa,

1451: ``The string landscape and the swampland,''

1452:  arXiv:hep-th/0509212.

1453:  %%CITATION = HEP-TH 0509212;%%

1454: 

1455: \bibitem{Arkani-Hamed:2006dz}

1456:  N.~Arkani-Hamed, L.~Motl, A.~Nicolis and C.~Vafa,

1457: ``The string landscape, black holes and gravity as the weakest force,''

1458:  arXiv:hep-th/0601001.

1459:  %%CITATION = HEP-TH 0601001;%%

1460: 

1461: \bibitem{Silverstein:2003hf}

1462:  E.~Silverstein and D.~Tong,

1463: ``Scalar speed limits and cosmology: Acceleration from D-cceleration,''

1464:  Phys.\ Rev.\ D {\bf 70}, 103505 (2004)

1465:  [arXiv:hep-th/0310221].

1466:  %%CITATION = HEP-TH 0310221;%%

1467: 

1468: \bibitem{Kofman:2004yc}

1469:  L.~Kofman, A.~Linde, X.~Liu, A.~Maloney, L.~McAllister and E.~Silverstein,

1470: ``Beauty is attractive: Moduli trapping at enhanced symmetry points,''

1471:  JHEP {\bf 0405}, 030 (2004)

1472:  [arXiv:hep-th/0403001].

1473:  %%CITATION = HEP-TH 0403001;%%

1474: 

1475: \bibitem{Watson:2004aq}

1476:  S.~Watson,

1477: ``Moduli stabilization with the string Higgs effect,''

1478:  Phys.\ Rev.\ D {\bf 70}, 066005 (2004)

1479:  [arXiv:hep-th/0404177].

1480:  %%CITATION = HEP-TH 0404177;%%

1481: 

1482: \bibitem{Alishahiha:2004eh}

1483:  M.~Alishahiha, E.~Silverstein and D.~Tong,

1484: ``DBI in the sky,''

1485:  Phys.\ Rev.\ D {\bf 70}, 123505 (2004)

1486:  [arXiv:hep-th/0404084].

1487:  %%CITATION = HEP-TH 0404084;%%

1488: 

1489: \bibitem{Brustein:2002xf}

1490:  R.~Brustein, S.~P.~de Alwis and E.~G.~Novak,

1491: ``M-theory moduli space and cosmology,''

1492:  Phys.\ Rev.\ D {\bf 68}, 043507 (2003)

1493:  [arXiv:hep-th/0212344].

1494:  %%CITATION = HEP-TH 0212344;%%

1495: 

1496: \bibitem{Dine:1998qr}

1497:  M.~Dine, Y.~Nir and Y.~Shadmi,

1498: ``Enhanced symmetries and the ground state of string theory,''

1499:  Phys.\ Lett.\ B {\bf 438}, 61 (1998)

1500:  [arXiv:hep-th/9806124].

1501:  %%CITATION = HEP-TH 9806124;%%

1502: 

1503: \bibitem{Russo:1996if}

1504:  J.~G.~Russo and A.~A.~Tseytlin,

1505: ``Waves, boosted branes and BPS states in M-theory,''

1506:  Nucl.\ Phys.\ B {\bf 490}, 121 (1997)

1507:  [arXiv:hep-th/9611047].

1508:  %%CITATION = HEP-TH 9611047;%%

1509: 

1510: \bibitem{Schwarz:1995du}

1511:  J.~H.~Schwarz,

1512: ``Superstring dualities,''

1513:  Nucl.\ Phys.\ Proc.\ Suppl.\  {\bf 49}, 183 (1996)

1514:  [arXiv:hep-th/9509148].

1515:  %%CITATION = HEP-TH 9509148;%%

1516: 

1517: \bibitem{Schwarz:1996bh}

1518:  J.~H.~Schwarz,

1519: ``Lectures on superstring and M theory dualities,''

1520:  Nucl.\ Phys.\ Proc.\ Suppl.\  {\bf 55B}, 1 (1997)

1521:  [arXiv:hep-th/9607201].

1522:  %%CITATION = HEP-TH 9607201;%%

1523: 

1524: \bibitem{Bergshoeff:1987qx}

1525:  E.~Bergshoeff, E.~Sezgin and P.~K.~Townsend,

1526: ``Properties Of The Eleven-Dimensional Super Membrane Theory,''

1527:  Annals Phys.\  {\bf 185}, 330 (1988).

1528:  %%CITATION = APNYA,185,330;%%

1529: 

1530: \bibitem{Duff:1987cs}

1531:  M.~J.~Duff, T.~Inami, C.~N.~Pope, E.~Sezgin and K.~S.~Stelle,

1532: ``Semiclassical Quantization Of The Supermembrane,''

1533:  Nucl.\ Phys.\ B {\bf 297}, 515 (1988).

1534:  %%CITATION = NUPHA,B297,515;%%

1535: 

1536: \bibitem{Russo:1996ph}

1537:  J.~G.~Russo,

1538: ``Supermembrane dynamics from multiple interacting strings,''

1539:  Nucl.\ Phys.\ B {\bf 492}, 205 (1997)

1540:  [arXiv:hep-th/9610018].

1541:  %%CITATION = HEP-TH 9610018;%%

1542: 

1543: \bibitem{Friess:2004zk}

1544:  J.~J.~Friess, S.~S.~Gubser and I.~Mitra,

1545: ``String creation in cosmologies with a varying dilaton,''

1546:  Nucl.\ Phys.\ B {\bf 689}, 243 (2004)

1547:  [arXiv:hep-th/0402156].

1548:  %%CITATION = HEP-TH 0402156;%%

1549: 

1550: \bibitem{Battefeld:2005av}

1551:  T.~Battefeld and S.~Watson,

1552: ``String gas cosmology,''

1553:  arXiv:hep-th/0510022.

1554:  %%CITATION = HEP-TH 0510022;%%

1555: 

1556: \bibitem{Brandenberger:2005nz}

1557:  R.~H.~Brandenberger,

1558: ``Challenges for string gas cosmology,''

1559:  arXiv:hep-th/0509099.

1560:  %%CITATION = HEP-TH 0509099;%%

1561: 

1562: \bibitem{Brandenberger:2005fb}

1563:  R.~H.~Brandenberger,

1564: ``Moduli stabilization in string gas cosmology,''

1565:  arXiv:hep-th/0509159.

1566:  %%CITATION = HEP-TH 0509159;%%

1567: 

1568: \bibitem{Patil:2005nm}

1569:   S.~P.~Patil,

1570:   ``Moduli (dilaton, volume and shape) stabilization via massless F and D

1571:   string modes,''

1572:   arXiv:hep-th/0504145.

1573:   %%CITATION = HEP-TH 0504145;%%

1574:   %%Cited 14 times in SPIRES-HEP

1575:   

1576: \bibitem{Kaloper:2004yj}

1577:  N.~Kaloper, J.~Rahmfeld and L.~Sorbo,

1578: ``Moduli entrapment with primordial black holes,''

1579:  Phys.\ Lett.\ B {\bf 606}, 234 (2005)

1580:  [arXiv:hep-th/0409226].

1581:  %%CITATION = HEP-TH 0409226;%%

1582: 

1583: \bibitem{Abel:2005jx}

1584:  S.~A.~Abel and J.~Gray,

1585: ``On the chaos of D-brane phase transitions,''

1586:  JHEP {\bf 0511}, 018 (2005)

1587:  [arXiv:hep-th/0504170].

1588:  %%CITATION = HEP-TH 0504170;%%

1589: 

1590: \bibitem{Mohaupt:2005pa}

1591:  T.~Mohaupt and F.~Saueressig,

1592: ``Conifold cosmologies in IIA string theory,''

1593:  Fortsch.\ Phys.\  {\bf 53}, 522 (2005)

1594:  [arXiv:hep-th/0501164].

1595:  %%CITATION = HEP-TH 0501164;%%

1596: 

1597: \bibitem{Berndsen:2005qq}

1598:  A.~Berndsen, T.~Biswas and J.~M.~Cline,

1599: ``Moduli stabilization in brane gas cosmology with superpotentials,''

1600:  JCAP {\bf 0508}, 012 (2005)

1601:  [arXiv:hep-th/0505151].

1602:  %%CITATION = HEP-TH 0505151;%%

1603: 

1604: \bibitem{Strominger:1995cz}

1605:  A.~Strominger,

1606: ``Massless black holes and conifolds in string theory,''

1607:  Nucl.\ Phys.\ B {\bf 451}, 96 (1995)

1608:  [arXiv:hep-th/9504090].

1609:  %%CITATION = HEP-TH 9504090;%%

1610: 

1611: \bibitem{Lidsey:1999mc}

1612:  J.~E.~Lidsey, D.~Wands and E.~J.~Copeland,

1613: ``Superstring cosmology,''

1614:  Phys.\ Rept.\  {\bf 337}, 343 (2000)

1615:  [arXiv:hep-th/9909061].

1616:  %%CITATION = HEP-TH 9909061;%%

1617: 

1618: \bibitem{Lawrence:1995ct}

1619:  A.~E.~Lawrence and E.~J.~Martinec,

1620: ``String field theory in curved spacetime and the resolution of spacelike

1621: singularities,''

1622:  Class.\ Quant.\ Grav.\  {\bf 13}, 63 (1996)

1623:  [arXiv:hep-th/9509149].

1624:  %%CITATION = HEP-TH 9509149;%%

1625: 

1626: \bibitem{Gubser:2003vk}

1627:  S.~S.~Gubser,

1628: ``String production at the level of effective field theory,''

1629:  Phys.\ Rev.\ D {\bf 69}, 123507 (2004)

1630:  [arXiv:hep-th/0305099].

1631:  %%CITATION = HEP-TH 0305099;%%

1632: 

1633: \bibitem{Birrell:1982ix}

1634:  N.~D.~Birrell and P.~C.~W.~Davies,

1635: ``Quantum Fields In Curved Space,''

1636: %\href{http://www.slac.stanford.edu/spires/find/hep/www?irn=998621}{SPIRES entry}

1637: 

1638: \bibitem{Chung:1998bt}

1639:  D.~J.~H.~Chung,

1640: ``Classical inflation field induced creation of superheavy dark matter,''

1641:  Phys.\ Rev.\ D {\bf 67}, 083514 (2003)

1642:  [arXiv:hep-ph/9809489].

1643:  %%CITATION = HEP-PH 9809489;%%

1644: 

1645: \bibitem{Watson:2003gf}

1646:  S.~Watson and R.~Brandenberger,

1647: ``Stabilization of extra dimensions at tree level,''

1648:  JCAP {\bf 0311}, 008 (2003)

1649:  [arXiv:hep-th/0307044].

1650:  %%CITATION = HEP-TH 0307044;%%

1651: 

1652: \end{thebibliography}

1653: 

1654: \end{document}

1655: 

1656: