hep-th0510129/TCT.tex
1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: %                                                                %
3: %   D-instantons and Closed String Tachyons                      %
4: %   in Misner Space                      (v.2)  2006/01/05       %
5: %                                                                %
6: %   Yasuaki Hikida and Ta-Sheng Tai                              %
7: %                                                                %
8: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
9: 
10: \documentclass[12pt,a4paper]{article}
11: \usepackage{type1cm,amsmath,hangcaption,graphicx,indentfirst}
12: %\usepackage{showkeys} %This should be commented out.
13: \usepackage[psamsfonts]{amssymb}
14: \numberwithin{equation}{section}
15: 
16: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
17: %                        page size
18: 
19: \setlength{\textwidth}{16truecm}
20: \setlength{\oddsidemargin}{0truecm}
21: \setlength{\evensidemargin}{\oddsidemargin}
22: \setlength{\textheight}{24truecm}
23: \setlength{\topmargin}{0truecm}
24: \setlength{\headheight}{0truecm}
25: \jot 3ex
26: 
27: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
28: %                          often used macro
29: 
30: \newcommand{\kett}[1]{{\left|#1\right\rangle}}
31: \newcommand{\braa}[1]{{\left\langle#1\right|}}
32: \newcommand{\ket}[1]{{|#1\rangle}{}}
33: \newcommand{\bra}[1]{{\langle#1|}}
34: \newcommand{\tq}{\tilde{q}}
35: \newcommand{\nn}{\nonumber\\}
36: \newcommand{\ts}[1]{{\textstyle #1}}
37: \DeclareMathOperator*{\Tr}{{\rm Tr}}
38: \DeclareMathOperator*{\re}{{\rm Re}}
39: \DeclareMathOperator*{\im}{{\rm Im}}
40: \DeclareMathOperator*{\arccosh}{{\rm arccosh}}
41: \DeclareMathOperator*{\arcsinh}{{\rm arcsinh}}
42: 
43: %\newcommand{\cleqn}{\setcounter{equation}{0}}
44: 
45: 
46: 
47: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
48: 
49: %\renewcommand{\baselinestretch}{1.3}
50: \renewcommand{\baselinestretch}{1.2}
51: 
52: \begin{document}
53: %%% Title page %%%%%
54: \begin{titlepage}
55: 
56:  \renewcommand{\thefootnote}{\fnsymbol{footnote}}
57: \begin{flushright}
58:  \begin{tabular}{l}
59:  KEK-TH-1043\\
60:  UT-05-15 \\
61:  hep-th/0510129\\
62: % \today %This should be commented out.
63:  \end{tabular}
64: \end{flushright}
65: 
66:  \vfill
67:  \begin{center}
68:  \font\titlerm=cmr10 scaled\magstep4
69:  \font\titlei=cmmi10 scaled\magstep4
70:  \font\titleis=cmmi7 scaled\magstep4
71:  \centerline{\titlerm D-instantons and Closed String Tachyons}
72:  \vskip .3 truecm
73:  \centerline{\titlerm in Misner Space}
74:  \vskip 2.5 truecm
75: 
76: \noindent{ \large Yasuaki Hikida$^a$\footnote{E-mail:
77: hikida@post.kek.jp} and Ta-Sheng Tai$^b$\footnote{E-mail:
78: tasheng@hep-th.phys.s.u-tokyo.ac.jp}}
79: \bigskip
80: 
81:  \vskip .6 truecm
82: \centerline{\it $^a$ Theory Group, High Energy Accelerator Research Organization (KEK)}
83: \centerline{\it Tukuba, Ibaraki 305-0801, Japan}
84: \bigskip
85: \centerline{\it $^b$ Department of Physics,  Faculty of Science, 
86: University of Tokyo}
87: \centerline{\it Hongo, Bunkyo-ku, Tokyo 113-0033, Japan}
88: 
89:  \vskip .4 truecm
90: 
91: 
92:  \end{center}
93: 
94:  \vfill
95: \vskip 0.5 truecm
96: 
97: 
98: \begin{abstract}
99: 
100: We investigate
101: closed string tachyon condensation
102: in Misner space, a toy model for 
103: big bang universe.
104: In Misner space, we are able to condense tachyonic 
105: modes of closed strings in the twisted sectors,
106: which is supposed to remove the big bang singularity.
107: In order to examine this, we utilize D-instanton as a probe. 
108: First, we study general properties of D-instanton
109: by constructing boundary state and 
110: effective action.
111: Then, resorting to these, we are able to show that tachyon condensation actually deforms the geometry
112: such that the singularity becomes milder.
113: 
114: \end{abstract}
115: \vfill
116: \vskip 0.5 truecm
117: 
118: \setcounter{footnote}{0}
119: \renewcommand{\thefootnote}{\arabic{footnote}}
120: \end{titlepage}
121: 
122: \newpage
123: 
124: \tableofcontents
125: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
126: %\newpage
127: \section{Introduction}
128: \label{Intoduction}
129: 
130: In general relativity, we may encounter singularities at
131: the early universe or inside black holes, and stringy effects
132: are supposed to resolve these singularities.
133: However, it is difficult to show this explicitly
134: since string theory on these backgrounds is not solvable
135: in general.
136: Therefore, it is useful to investigate simpler models for 
137: cosmological backgrounds by utilizing Lorentzian orbifolds as in
138: \cite{KOSST,BHKN,CC,Nekrasov,Simon,LMS1,CCK,Lawrence,LMS2,FM,HP}
139: or gauged WZW models as in \cite{NW,EGKR,CKR,HT,TT}.
140: In this paper, we consider a simplest model for the early
141: universe, namely Misner (Milne) space \cite{Misner}.
142: Strings in this background are investigated in 
143: \cite{Nekrasov,BCKR,Pioline1,Pioline2,Pioline3,Pioline4}
144: and D-branes are in \cite{HNP2}.%
145: \footnote{See, e.g., \cite{HS,AP,DN,CLO,Okuyama,HNP} for D-branes in 
146: other time-dependent backgrounds.}
147: 
148: Misner space can be constructed as (1+1) dimensional
149: Minkowski space with the identification of a discrete boost.
150: This space consists of four regions, which are classified into
151: two types. One type are called as cosmological regions,
152: which include big bang/big crunch singularity, and 
153: the other are called as whisker regions, 
154: which include closed time-like curves.
155: In bosonic string or superstring theories with opposite
156: periodic condition of space-time fermions,  
157: there are tachyonic modes
158: in the twisted closed string sector. We find that
159: there are two types of tachyonic modes, and one of them
160: is localized in the whisker regions.
161: 
162: It is therefore natural to expect that the condensation
163: of the tachyon removes the regions with closed time-like curves \cite{Hagedorn}. 
164: Furthermore, it still remains crucial 
165: if the condensation affects the cosmological regions 
166: through the big bang singularity 
167: and ultimately resolves the singularity \cite{MS} 
168: (see also \cite{Silverstein}).
169: This is an analogous situation to the condensation of
170: localized tachyon in Euclidean orbifolds, say, 
171: ${\mathbb C}/{\mathbb Z}_N$ \cite{APS}.
172: See also \cite{Martinec,HMT} and references therein.
173: 
174: In order to investigate the tachyon condensation,
175: we utilize D-instantons in the Misner space.%
176: \footnote{While completing this work, a paper analyzing 
177: similar situation appeared in the arXiv \cite{She},
178: where the analysis in null brane case \cite{BKRS} was 
179: directly applied.
180: The relation to matrix models has been discussed 
181: recently in \cite{CSV,Li,LS,pp,Chen,RS} as well.}
182: We find that there are D-instantons localized in the
183: big crunch/big bang singularity, which are analogous
184: to fractional branes in orbifold models \cite{DDG,DG,BCR}.
185: Due to this analogy, we call this type of instantons as 
186: ``fractional'' instantons.
187: By summing every fractional instantons, we can construct
188: a type of D-instanton away from the big crunch/big bang
189: singularity. We construct effective theory for open strings
190: on this instanton, and probe the background geometry 
191: using its moduli space.
192: Before condensing the tachyon, 
193: the geometry read off from the moduli is Misner space
194: as expected. After turning on a small tachyon vev,
195: we find the singularity becomes milder than the original
196: background, which is consistent with the conclusion in \cite{MS}.
197: Concretely, we find a flow from Misner space with smaller space cycle
198: to larger space cycle and a flow from Misner space to the future patch
199: of Minkowski space in Rindler coordinates.
200: 
201: 
202: The organization of this paper is as follows.
203: First, we introduce Misner space as a Lorentzian orbifold,
204: and study closed strings both in untwisted and twisted
205: sectors. In particular, we closely investigate the
206: properties of the tachyonic modes of closed strings
207: in twisted sectors, and observe one type of the tachyon
208: is localized in the whisker regions.
209: In order to see how the condensation of the tachyonic modes
210: changes the background geometry, we utilize D-instantons
211: since they are suitable to probe the deformed geometry \cite{APS,BKRS}.
212: Using the boundary state formalism, we confirm the existence
213: of fractional instantons.
214: We construct effective theory for open strings on the sum
215: of every fractional instantons, and probe the geometry following 
216: \cite{APS}.
217: {}From this probe, we see that the deformation by the condensation
218: of the closed string tachyon tends to make the big bang singularity milder.
219: 
220: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
221: \section{Closed string tachyons in Misner space}
222: \label{tachyon}
223: 
224: 
225: For string theory application, it is convenient to
226: define Misner space \cite{Misner} as
227: (1+1) dimensional Minkowski space with
228: identification through Lorentz boost.
229: We denote the light-cone coordinates as
230: $x^{\pm} = \frac{1}{\sqrt2} (x^0 \pm x^1)$,
231: then the Misner space is defined by using the identification
232: \begin{align}
233: x^{\pm} &\sim g \cdot x^{\pm} ~,
234: & g \cdot x^{\pm} &= e^{\pm 2 \pi \gamma} x^{\pm} ~,
235: \label{misner}
236: \end{align}
237: where  $\gamma \in {\mathbb R}$.
238: 
239: The space is divided into four regions by lines
240: $x^+ x^- = 0$.
241: We are mainly interested in $x^+ , x^- > 0$ region,
242: which can be regarded as a big bang universe.
243: In this region, we change the coordinate by
244: $x^{\pm} = \frac{1}{\sqrt{2}}t e^{\pm \psi}$,
245: then the metric is
246: \begin{align}
247: ds^2 = -dt^2 + t^2 d \psi ^2
248: \label{cosmologicalr}
249: \end{align}
250: with the periodicity $\psi \sim \psi + 2 \pi \gamma$.
251: The radius of the space cycle depends on the time $t$,
252: and the radius vanishes when $t=0$. Thus, the space
253: starts at the big bang and expands as time goes.
254: The region with $x^+ , x^- < 0$ is the time reversal
255: of the big bang universe, namely a big crunch universe.
256: The metric is also given by \eqref{cosmologicalr}.
257: The space starts at $t=-\infty$ with infinitely large
258: radius, shrinks as time goes, and meets the big
259: crunch when $t=0$. We can deal with this region
260: in a way similar to the big bang region by
261: time reversal. 
262: 
263: 
264: The regions with $x^+ x^- > 0$ are called as 
265: cosmological regions, and the other two
266: regions with  $x^+ x^- < 0$ are referred to
267: whisker regions.
268: Under the coordinate transformation
269: $x^{\pm} = \pm  \frac{1}{\sqrt{2}} r e^{\pm \chi}$,
270: the metric in the whisker regions is given by
271: \begin{align}
272: ds^2 = d r^2 - r^2 d \chi^2 ~.
273: \label{whiskerr}
274: \end{align}
275: There are closed time-like curves everywhere in these regions,
276: which are attributed to 
277: the periodicity $\chi \sim \chi + 2 \pi \gamma$ of time coordinate. 
278: We will not study the whisker regions in detail,
279: but it was suggested in $\cite{Hagedorn}$ that these regions might be removed from the space-time
280: by tachyon condensation.
281: 
282: 
283: In addition to Misner space, $d$ extra flat directions are included to form critical string theory, 
284: where $d=24$ for bosonic case and $d=8$ for superstring case. In this section, 
285: the untwisted sector and the twisted sectors for bosonic case are analyzed at first, and superstring cases follow soon.
286: Investigation of the behavior of tachyonic modes is of main interest,%
287: \footnote{Similar discussion was made in \cite{Hagedorn}
288: for tachyonic modes in O-plane.}
289: whose condensation will be considered below.
290: 
291: 
292: \subsection{Closed strings in the untwisted sector}
293: 
294: 
295: Let us first examine the untwisted sector of closed strings.
296: To have a quick grasp of the behavior of the closed strings, 
297: it is sufficient to focus on the zero mode part, nevertheless higher modes can be included easily.
298: The untwisted sector is obtained by summing all images
299: of the discrete boost \eqref{misner} in the covering space,
300: that is, two dimensional Minkowski space-time.
301: The zero mode part reduces to  a particle in the untwisted sector,
302: and its trajectory in the covering space is
303: a straight line
304: \begin{align}
305:  X^{\pm} = x_0^{\pm} + \alpha ' p^{\pm} \tau ~.
306: \end{align}
307: The mass square is given by $m^2 = 2 p^+ p^-$, which can be 
308: positive or negative in the bosonic string or type 0 superstring 
309: theory. For massive modes $p^+, p^- > 0$,\footnote{The particle
310: with $p^+, p^- < 0$ travels from future to past, or it can
311: be considered as an anti-particle.}
312: the particle starts from infinite past in the big crunch region
313: and ends at infinite future in the big bang region.
314: While cruising over the space-time, the particle crosses a whisker
315: region only up to a point $| r_0 | = |l|/m$ with 
316: $l=x^+_0 p^- - x^-_0 p^+$.
317: For the tachyonic modes, we have to set $p^+ < 0,p^- > 0$
318: or  $p^+ > 0,p^- < 0$.
319: The trajectory begins from spatial infinity in a whisker region, 
320: crosses a cosmological region and terminates at the spatial 
321: infinity in the other whisker region. 
322: In other words, the tachyonic modes exist
323: only for a short time $0 < t < t_0=|l|/\sqrt{- m^2}$ 
324: or $0 > t > t_0= -|l|/\sqrt{- m^2}$ in 
325: cosmological regions.
326: 
327: 
328: When it comes to the bosonic string case, the Virasoro condition reads
329: \begin{align}
330: L_0 + \tilde L_0 -2 & = \frac{\alpha '}{2} ( - 2 p^+ p^- + m^2)  = 0 ~,
331: &m^2 &= \vec k ^2 + \frac{2}{\alpha '} (N + \tilde N - 2) ~,
332: \end{align}
333: where $\vec k$ represents the momenta along the extra
334: directions and $N,\tilde N$ denote occupation numbers.
335: For our analysis, it is convenient to rewrite as
336: \begin{align}
337:  p_{\eta}^2  + V(\eta) &=  0~,
338: &p_{\phi}^2 + V(\phi) &= 0 ~,
339: \label{dis}
340: \end{align}
341: where
342: \begin{align}
343: V(\eta) &= - l^2 - m^2 e^{2 \eta}  ~,
344: &V(\phi) &= - l^2 + m^2 e^{2 \phi} ~.
345: \label{potential}
346: \end{align}
347: We have changed the coordinates as $t = \pm e^{\eta},r = \pm e^{\phi}$.
348: For massive modes $m^2 > 0$, the condition $p^2_{\eta} \geq 0$ is 
349: always satisfied, thus the string runs from infinite past to 
350: infinite future. 
351: On the other hand, the condition $p^2_{\phi} \geq 0$ restricts
352: $|r| \leq |l|/m$.
353: For tachyonic modes $m^2 < 0$, the condition $p^2_{\phi} \geq 0$
354: is always satisfied, while the condition $p^2_{\eta} \geq 0$
355: restricts $|t| \leq |l|/\sqrt{- m^2}$.
356: These results are consistent with the previous analysis.
357: 
358: For a scalar particle, Klein-Gordon equation can be
359: obtained by replacing $p_{\eta},p_{\phi}$
360: with derivatives $i\partial_{\eta},i\partial_{\phi}$.
361: The wave function shows oscillatory or damping behavior
362: above or below the potential.
363: This implies that the wave function is localized in the cosmological
364: regions for massive modes and in the whisker regions for
365: tachyon modes. In fact, the wave function can be expressed 
366: in the covering space as \cite{Nekrasov}
367: \begin{align}
368: \Psi_{p,l} = \int d w e^{i(p^+ X^- e^{-\gamma w}+ p^- X^+ e^{\gamma w} + i w l + \vec k \cdot \vec X)} ~,
369: \end{align}
370: which is invariant under the orbifold action.
371: For massive modes, we can set $p^+ = p^- = m/\sqrt2 $,
372: then the wave function is localized in the cosmological regions.
373: In particular, for large $ - X^+X^- \gg |l|/m$, the wave function 
374: shows damping behavior \cite{Nekrasov} as 
375: $\Psi_{p,l} \propto \exp (- m r)$.
376: For tachyonic case, we may set 
377: $p^+ = - p^- = \sqrt{- m^2 /2 }$ or
378: $p^+ = - p^- = -\sqrt{- m^2 /2 } $.
379: Then, the wave function is localized in the whisker regions
380: and shows damping behavior in the cosmological regions as
381: $\Psi_{p,l} \propto \exp (- \sqrt{- m^2} |t|)$ for 
382: $X^+ X^- \gg |l|/\sqrt{-m^2}$.
383: 
384: 
385: \subsection{Closed strings in the twisted sectors}
386: 
387: 
388: In the orbifold theory, there are twisted sectors
389: of closed strings. Because of the identification \eqref{misner}, twisted periodic conditions give rise to
390: \begin{align}
391:  X^{\pm} (\tau, \sigma+ 2\pi) = e^{\pm 2\pi
392: \gamma w} X^{\pm}(\tau, \sigma)
393: \label{twist}
394: \end{align}
395: with non-trivial $w \in {\mathbb Z}$. 
396: There are two types of twisted closed strings, 
397: since two types of cycles, on which closed strings
398: can be wrapped, namely, $\psi$-cycle in the cosmological
399: regions and $\chi$-cycle in the whisker regions.
400: 
401: 
402: 
403: The lowest mode of the
404: solution to \eqref{twist} is
405: \begin{align}
406: X^{\pm} (\tau, \sigma) &= \pm \sqrt{\frac{\alpha '}{2}} \frac{\alpha^{\pm}_0}{\nu} e^{\pm \nu (\tau + \sigma) }
407:  \mp \sqrt{\frac{\alpha '}{2}} \frac{\tilde \alpha^{\pm}_0}{\nu} e^{\mp \nu (\tau - \sigma) } ~,
408:  & \nu &= \gamma w 
409:  \label{modes}
410: \end{align}
411: which would provide clear physical picture of the winding strings.
412: For bosonic strings, the Virasoro constraints read
413: \begin{align}
414: \omega^2 &= \frac{1}{2}(\alpha^+_0 \alpha^-_0 + \alpha^-_0 \alpha^+_0) ~,
415: & \tilde \omega^2
416: = \frac12( \tilde \alpha^+_0 \tilde \alpha^-_0 + \tilde \alpha^-_0\tilde
417: \alpha^+_0)
418: \end{align}
419: with
420: \begin{align}
421: \omega^2 &= \frac{\alpha '}{4} \vec k ^2 + N - 1 + \frac12 \nu^2 ~,
422: &\tilde \omega^2 &= \frac{\alpha '}{4} \vec k ^2 + \tilde N - 1 + \frac12 \nu^2 ~.
423: \label{tomega}
424: \end{align}
425: {}From the above equations, we can see that 
426: $\omega ^2 , \tilde \omega ^2$ can take positive and negative
427: values for small $ | \nu | < \sqrt2$.
428: As shown below, there is no shift of zero point energy 
429: $\frac12 \nu^2$ in superstring theories.
430: Under the lowest mode expansion,
431: the level matching condition demands
432: $\omega ^2 = \tilde \omega ^2$, 
433: thus there is no angular momentum 
434: $\omega ^2 - \tilde \omega ^2 = \nu l=0$.
435: In fact, it was shown in \cite{HNP2} that
436: the general on-shell states are
437: isomorphic to the states with only lowest modes
438: in Misner space part.
439: 
440: 
441: 
442: 
443: For this reason, we will only consider the case with $l=0$ in the following.
444: For massive case $\omega^2 >0$,
445: we can set \cite{Pioline1}
446: \begin{align}
447:  \alpha^+_0 &= \alpha^-_0 = \epsilon \omega  ~,
448:  &\tilde \alpha^+_0 &= \tilde \alpha^-_0 = \tilde \epsilon \omega  ~,
449:  &\epsilon,\tilde \epsilon &= \pm 1
450: \end{align}
451: without loss of generality.
452: With these solutions, we obtain
453: \begin{align}
454: X^{\pm} (\tau,\sigma) =
455:  \frac{\sqrt{2 \alpha '}\omega}{\nu} e^{\pm \nu \sigma} \sinh \nu \tau
456:  \label{short}
457: \end{align}
458: for $\epsilon = \tilde \epsilon = 1$.
459: This represents a string wrapping the whole cosmological regions,
460: namely, winding $\psi$-cycle.
461: Another choice may be
462: $\epsilon = - \tilde \epsilon = 1$, which leads to
463: \begin{align}
464: X^{\pm} (\tau,\sigma ) =
465: \pm \frac{\sqrt{2 \alpha '}\omega}{\nu} e^{\pm \nu \sigma} \cosh \nu \tau ~.
466:  \label{long}
467: \end{align}
468: This string winds $\chi$-cycle in the whisker regions,
469: and exists from $r_0= 2 \sqrt{\alpha '} \omega / \nu$ to spatial infinity.
470: See fig. \ref{winding} for these classical trajectories.
471: \begin{figure}
472: \centerline{\scalebox{0.6}{\includegraphics{winding.eps}}}
473: \caption{\it Winding strings with no angular momentum $l=0$.
474: (A) For massive case,
475: winding strings are wrapped on the whole cosmological regions
476: or a part of whisker regions from finite $r_0$ to spatial infinity. 
477: (B) For tachyonic case,
478: winding strings are wrapped on the whole whisker regions
479: or a part of cosmological regions from finite $t_0$ to past or future 
480: infinity.}
481: \label{winding}
482: \end{figure}
483: 
484: For tachyonic case $\omega^2 =  - {\cal M}^2 < 0$,
485: we can set 
486: \begin{align}
487:  \alpha^+_0 &= - \alpha^-_0 = \epsilon {\cal M}  ~,
488:  &\tilde \alpha^+_0 &= - \tilde \alpha^-_0 = \tilde \epsilon {\cal M}  ~,
489:  &\epsilon,\tilde \epsilon &= \pm 1 ~.
490: \end{align}
491: When $\epsilon = \tilde \epsilon = 1$,
492: the classical trajectory is 
493: \begin{align}
494: X^{\pm} (\tau,\sigma) = \pm
495:  \frac{\sqrt{2\alpha'} {\cal M}}{\nu} e^{\pm \nu \sigma} \sinh \nu \tau ~,
496: \end{align}
497: which wraps on the whole whisker region.
498: A good property for us is that the tachyon touches the cosmological
499: regions only through the big crunch/big bang singularity.
500: This is analogous to the tachyons in twisted sectors of Euclidean 
501: orbifolds, which are localized at their fixed points.
502: Using this fact, the authors of \cite{APS} can discuss closed
503: tachyon condensation, since the condensation affects only
504: small regions of backgrounds.
505: Therefore, it is natural to think that we can also
506: analyze the closed tachyon condensation in the cosmological regions, 
507: even though it seems difficult to know the fate of the whisker 
508: regions.%
509: \footnote{It was suggested in \cite{Hagedorn} that the tachyon 
510: condensation removes the whisker regions from the background.}
511: When $\epsilon = - \tilde \epsilon = 1$,
512: the classical trajectory is 
513: \begin{align}
514: X^{\pm} (\tau,\sigma ) =
515: \frac{\sqrt{2\alpha'} {\cal M}}{\nu} e^{\pm \nu \sigma} \cosh \nu \tau ~,
516: \end{align}
517: which stands for the process of string pair creation
518: at a time $t_0 = 2 \sqrt{\alpha '} {\cal M}/\nu $ 
519: and ends at future infinity.%
520: \footnote{We assumed ${\cal M}/\nu > 0$. If ${\cal M}/\nu < 0$, 
521: then the trajectory represents the time-reversal process.}
522: We will not consider the condensation of this tachyon
523: because it has nothing to do with the resolution of big crunch/big bang
524: singularity.
525: 
526: 
527: In quantum level, the modes are treated as operators
528: with commutation relations
529: \begin{align}
530:  [ \alpha^+_0 , \alpha^-_0 ] &= - i \nu ~,
531: &[ \tilde \alpha^+_0 , \tilde \alpha^-_0 ] &= i \nu ~.
532: \end{align}
533: Following \cite{Pioline1}, we use a representation
534: \begin{align}
535:  \alpha ^{\pm}_0 &= i \sqrt{\frac{\alpha '}{2}} \partial_{\mp} \pm 
536:       \frac{\nu} {\sqrt{2 \alpha '}}  x^{\pm} ~,
537:  &\tilde \alpha ^{\pm}_0 &= 
538:  i \sqrt{\frac{\alpha '}{2}} \partial_{\mp} \mp 
539:   \frac{\nu}{\sqrt{2 \alpha '}} x^{\pm} ~.
540: \end{align}
541: Then, the Virasoro conditions \eqref{tomega} reduce
542: to the equations \eqref{dis}
543: with the potentials \cite{Pioline1}
544: \begin{align}
545: V(\eta) &= - \frac{\nu ^2 e^{4 \eta} }{ {\alpha '}^2} 
546:            - \frac{4 \omega ^2}{\alpha '} e^{2 \eta}  ~,
547: &V(\phi) &= -\frac{ \nu ^2 e^{4 \phi}}{ {\alpha '}^2}  
548:            + \frac{4 \omega^2}{\alpha '} e^{2 \phi} ~.
549: \end{align}
550: For massive case  $\omega ^2 > 0$, $p ^2 _{\eta} \geq 0$ does not
551: give any condition, but $p^2_{\phi} \geq 0$ is equivalent to
552: $| r | \geq 2 \sqrt{\alpha '} | \omega/\nu | $.
553: For tachyonic case  $\omega ^2 = - {\cal M}^2 < 0$,
554: $p^2_{\phi} \geq 0$ does not give any condition, but 
555: $p ^2 _{\eta} \geq 0$ is equivalent to 
556: $|t| \geq 2 \sqrt{\alpha '} | {\cal M}/\nu |$.
557: These conditions are the same as in the previous analysis.
558: Replacing $p_{\eta},p_{\phi}$ with derivatives 
559: $i \partial_{\eta}, i \partial_{\phi}$,
560: the Virasoro conditions lead to Klein-Gordon equation.
561: The solution should become oscillatory above the potential
562: and damping below the potential as mentioned before.
563: In each of the regions the wave functions can
564: be written in terms of Whittaker functions \cite{GS,Pioline1},
565: and in the whole space wave functions can be 
566: obtained by connecting those in the each region.
567: 
568: \subsection{Closed superstrings}
569: 
570: In this subsection, tachyonic modes 
571: in superstring cases are brought about 
572: by introducing worldsheet fermions and taking a particular GSO projections.
573: Due to the boundary condition \eqref{twist},
574: the full mode expansion for $X^{\pm}(\tau,\sigma)$ 
575: in $w$-th twisted sector is of the form
576: \begin{align}
577:  X^{\pm} (\tau,\sigma) = i \sqrt{\frac{\alpha '}{2}} 
578:   \sum_{n \in {\mathbb Z}} \left[
579:   \frac{\alpha^{\pm}_n}{n \pm i\nu} e^{-i(n \pm i\nu)(\tau + \sigma)} +
580:   \frac{\tilde \alpha^{\pm}_n}{n \mp i\nu} e^{-i(n \mp i\nu)(\tau - \sigma)}
581:   \right] ~,
582: \end{align}
583: where the oscillators satisfy the following commutation relations
584: \begin{align}
585:  [\alpha ^+_m,\alpha ^-_n] &= (- m - i \nu )\delta_{m+n} ~,
586: &[\tilde \alpha ^+_m, \tilde \alpha ^-_n] &= (- m + i \nu )\delta_{m+n} ~.
587: \end{align}
588: The boundary conditions for worldsheet fermions are
589: \begin{align}
590:  \psi^{\pm} (\tau , \sigma + 2 \pi)
591: &= - (-1)^A e^{\pm 2 \pi \nu} \psi^{\pm} (\tau , \sigma) ~,
592: &\tilde \psi^{\pm} (\tau , \sigma + 2 \pi)
593: &= - (-1)^A e^{\pm 2 \pi \nu} \tilde \psi^{\pm} (\tau , \sigma) ~,
594: \label{NSbc}
595: \end{align}
596: where $A=0$ for NS-sector and $A=1$ for R-sector.
597: The mode expansions can be given then
598: \begin{align}
599:  \psi^{\pm} &= \sum_{r \in  {\mathbb Z} + (A+1)/2}
600:   b^{\pm}_{-r} e^{-i(r \pm i \nu)(\tau + \sigma)} ~,
601:  &\tilde \psi^{\pm} &= \sum_{r \in  {\mathbb Z} + (A+1)/2}
602:   \tilde b^{\pm}_{-r} e^{-i(r \mp i \nu)(\tau - \sigma)} ~,
603: \end{align}
604: where the oscillators satisfy the anti-commutation relations
605: \begin{align}
606:  \{ b^{\pm}_r , b^{\mp}_s \} &= - \delta_{r+s} ~,
607:  &\{ \tilde b^{\pm}_r , \tilde b^{\mp}_s \} &= - \delta_{r+s} ~.
608: \end{align}
609: 
610: 
611: Combined with the bosonic part, the Virasoro constraint leads to%
612: \footnote{The counter part of right-mover is given in the same way.}
613: \begin{align}
614: \omega ^2 = \frac{\alpha '}{4} \vec k ^2 
615:  + N^{\pm}_{b} + N^{\pm}_{f} + N_{b} + N_{f} - \frac{1-A}{2} ~,
616:  \label{fvirasoro}
617: \end{align}
618: where $N_b,N_f$ are the occupation numbers of bosonic and
619: fermionic modes in extra $8$ directions.
620: Moreover,  $N^{\pm}_b,N^{\pm}_f$ are the occupation numbers of 
621: bosonic and fermionic modes of Misner space part.
622: They are explicitly given as
623: \begin{align}
624:  N^{\pm}_b = - \sum_{n > 0} \alpha^+_{- n} \alpha^-_{n}
625:              - \sum_{n > 0} \alpha^-_{- n} \alpha^+_{n}
626: \end{align}
627: for bosonic part and
628: \begin{align}
629: N^\pm_{f} = c(A) - \sum_{r > 0} (r - i \nu) b^+_{-r} b^-_r
630:    - \sum_{r > 0} (r + i \nu) b^-_{-r} b^+_r 
631: \end{align}
632: with $c(0) = 0$ and $c(1)= \frac{i \nu}{2} [ b_0^+ , b_0^- ]$
633: for fermionic part.
634: Notice that there is no zero energy shift in \eqref{fvirasoro}.
635: 
636: There appears to be tachyonic mode in NSNS sector, 
637: though it is projected out by the usual GSO projection. 
638: Nevertheless, it is possible to consistently impose 
639: usual GSO projection to even $w$-th sectors and
640: opposite GSO projection to odd $w$-th sectors.
641: More explicitly, we can use
642: \begin{align}
643: \begin{aligned}
644:  P^{\text {GSO} }_{\text {NSNS}} &= \frac14 \left(1 + (-1)^w(-1)^F \right)
645:                     \left(1 + (-1)^w(-1)^{\tilde F} \right) ~, \\
646:  P^{\text {GSO} }_{\text {RR}} &= \frac14 \left(1 + (-1)^w(-1)^F \right)
647:                     \left(1 \mp (-1)^w(-1)^{\tilde F} \right) ~,
648:                     \label{opposite}
649: \end{aligned}
650: \end{align}
651: where $F,\tilde F$ are the fermion number operators for
652: left and right moving parts, respectively. The sign $-$
653: is for type IIA and $+$ is for type IIB. 
654: The GSO projection for NSR or RNS sectors is defined 
655: in a similar manner.
656: Though the bulk tachyon is thrown away by this GSO projection,
657: in odd $w$-th twisted sectors tachyon mode survives.   
658: 
659: 
660: The partition function with this GSO projection can be
661: evaluated as \cite{Nekrasov,HNP2}\footnote{%
662: In order to compute partition functions or construct boundary
663: states, it is convenient to adopt a different way to construct
664: Hilbert space. Namely, we regard $\alpha^+_0,\alpha^-_0$
665: as creation and annihilation operators. For more detail,
666: see \cite{Pioline1,HNP2}.}
667: \begin{align}
668:  \int \frac{d^2 \tau}{16 \pi^2 \tau_2^2 \alpha '}
669:   \sum_{k,w}
670:   \frac{|\vartheta_3 (y|\tau) \vartheta_3 (0|\tau)^3
671:         - (-1)^w \vartheta_4 (y|\tau) \vartheta_4 (0|\tau)^3
672:         - (-1)^k \vartheta_2 (y|\tau) \vartheta_2 (0|\tau)^3
673:         |^2}{(4 \pi^2 \tau_2 \alpha ')^3
674:              |\vartheta_1 (y|\tau ) \eta (\tau )^9|^2}
675: \end{align}
676: with $y=i \gamma ( w \tau + k )$. 
677: The Hilbert spaces of all $w$-th twisted sector were summed over, 
678: and the projection operator $P = \sum_k g^k$ was inserted.
679: The factor $(-1)^k$ is needed to ensure the modular invariance,
680: and it means that space-time fermions should
681: give factor $-(-1)^k$ when going around $\tau_2$-cycle.
682: In other words, the GSO projection \eqref{opposite}
683: is equivalent to the requirement that space-time fermions
684: should transform under the orbifold action, such that, $g^k \cdot S^{\pm} = (-1)^k e^{\pm k \pi \gamma} S^{\pm}$.
685: Note that this action is of the same form as in \cite{APS}.
686: 
687: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
688: 
689: \section{D-instanton probe of closed tachyon condensation}
690: 
691: As shown in the previous section, there are tachyonic modes in the twisted sectors in bosonic string or superstring theories
692: with our choice of GSO projection.
693: We will focus on the big bang region, where the 
694: tachyonic modes are localized at the big bang singularity.
695: It is expected that the condensation of the tachyon
696: alters the geometry near the big bang singularity
697: just like the tip of cone of Euclidean orbifold is
698: deformed by localized closed string tachyon \cite{APS}.
699: 
700: In order to examine the deformed geometry,
701: D-instantons are shown to be useful as probes
702: for their promised properties.
703: We will see that a sum of ``fractional'' instantons enable us 
704: to look into what happens after the tachyon condensation. 
705: By means of the effective action on
706: it, we observe that the geometry is deformed
707: and manage to answer whether the tachyon
708: condensation resolves the singularity.
709: 
710: 
711: \subsection{Open strings on instantons and boundary states}
712: \label{bs}
713: 
714: One way to construct instantons in Misner space is
715: to sum over all image instantons in the covering space.
716: Namely, if we put an instanton at 
717: $x^{\pm} = x^{\pm}_0$, then we have to sum
718: up instantons at 
719: $x^{\pm} = e^{\pm 2 \pi \gamma k} x^{\pm}_0$
720: for all $k \in {\mathbb Z}$.
721: We will call this type as ``bulk'' instanton.
722: In the covering space, these image instantons are
723: the same as in the flat space case, thus the
724: annulus amplitude for open strings between
725: image instantons and boundary states for each
726: image instantons are the same as in the flat space case.
727: 
728: 
729: Interestingly, we can build another type of instantons
730: invariant under the orbifold action.
731: They are localized at the big crunch/big bang singularity
732: $x^{\pm} = 0$, and analogous to so-called fractional branes
733: in Euclidean orbifolds \cite{DDG,DG,BCR}.
734: We will call this type as ``fractional'' instanton even though
735: they are not fractional with respect to bulk instanton in any sense.
736: Let us start from the bulk instanton at the singularity,
737: which is the sum of all $i$-th image instantons in the covering space.
738: The orbifold action $g$ acts both on Chan-Paton 
739: indices and the states as \cite{DM}
740: \begin{align}
741:  g :| \psi , i , j \rangle 
742:  \rightarrow \sum_{i',j'= - \infty}^{\infty}
743:  \gamma_{i,i'} | \hat g \psi , i' , j' \rangle \gamma_{j' , j}^{-1} ~.
744:  \label{CP0}
745: \end{align}
746: We set $\gamma_{i,i '} = \delta_{i+1,i '}$ because
747: the orbifold action shifts $i$-th image brane into
748: $(i-1)$-th image brane.
749: The orbifold action to the states is represented by $\hat g$.
750: 
751: 
752: Another basis for Chan-Paton indices, that fits more into our practice, is%
753: \footnote{We neglect an overall factor $1/N$ with $N=\sum_i 1$.}
754: \begin{align}
755:  | \psi , m , n \rangle  
756:  = \sum_{i,j= - \infty}^{\infty} 
757: \xi^{im}| \psi , i , j \rangle \xi^{-jn}
758: \label{basis}
759: \end{align}
760: with a complex number $\xi$.
761: This is more adequate since Chan-Paton indices are invariant
762: under the orbifold action as
763: \begin{align}
764:  g :| \psi , m , n \rangle 
765:  \rightarrow 
766:  \xi^{m - n} | \hat g \psi , m , n \rangle  ~.
767: \label{CP}
768: \end{align}
769: Thanks to the invariance of this basis,
770: the instantons, between which open strings are stretched, are
771: invariant under the orbifold, and they in fact differ from the 
772: bulk instanton as will be explained.
773: If we are dealing with the orbifold ${\mathbb C}/{\mathbb Z}_N$,
774: then  $g^N = 1$ should be satisfied for all $(m,n)$,
775: which leads to $\xi = e^{2 \pi i /N}$ and 
776: $(m,n) \in {\mathbb Z}_N$.
777: This suggests that there are $N$ types of fractional branes,
778: which transform differently under the orbifold action. 
779: Since there is no such a periodicity condition in our case,
780: an arbitrary $\xi$ can be used.
781: Later we fix $\xi$ such that the effective theory includes
782: desired fields.
783: 
784: 
785: 
786: To extract physical spectrum of open strings with Chan-Paton
787: indices $(m,n)$, we project into invariant subspace 
788: under the orbifold action by
789: \begin{align}
790:  P &= \sum_{k=-\infty}^{\infty} g^k ~, 
791:    &g : X^{\pm} \to e^{\pm 2 \pi \gamma } X^{\pm} ~.
792: \end{align}
793: Since the open strings on instantons satisfy the Dirichlet
794: boundary condition
795: \begin{align}
796:  X^{\pm} (\tau , \sigma = 0 ) &=0 ~,
797: &X^{\pm} (\tau , \sigma = \pi) &=0 ~,
798: \end{align}
799: the mode expansion is give by
800: \begin{align}
801:  X^{\pm} (\tau , \sigma ) =  \sqrt{2 \alpha ' } 
802:  \sum_{n \neq 0} \frac{\alpha^{\pm}_n}{n} 
803:  e^{in \tau} \sin (n \sigma ) ~.
804: \end{align}
805: In terms of oscillators, the Virasoro generator and  the
806: twist operator are written as
807: \begin{align}
808:  L_0 &=
809:  - \sum_{n \geq 1} \alpha^+_{-n} \alpha^-_n
810:  - \sum_{n \geq 1} \alpha^-_{-n} \alpha^+_n ~,
811:  &g &= e^{2 \pi \gamma i \hat J } ~,
812:  &i\hat J  &=  - 
813:   \sum_{n > 0} 
814:  \left( \frac{\alpha^{+}_{-n} \alpha^{-}_{n}}{n} 
815:    - \frac{\alpha^{-}_{-n} \alpha^{+}_{n}}{n} \right) ~.
816: \end{align}
817: Gathering all of them,
818: the one-loop amplitude of the open strings can be computed as%
819: \footnote{We implicitly neglect $k=0$ sector since this part
820: is the same as the bulk brane case.}
821: \begin{align}
822:  Z_{mn} (it) = \sum_{k= -\infty}^{\infty}
823:   {\rm Tr}_{{\cal H}_{mn}} g^k e^{-2 \pi t (L_0 - \frac{1}{12})}
824:   = \sum_{k= -\infty}^{\infty}\xi ^{k(m-n)}
825:     \frac{2 i \sinh (\pi |k| \gamma )\eta (it)}{\vartheta_1 (i |k|\gamma |it) }
826:  ~.
827:   \label{parto}
828: \end{align}
829: Recall that 
830: the orbifold action also acts on the Chan-Paton indices 
831: as in \eqref{CP}.
832: Using the modular transformation $t \rightarrow s=1/t$,
833: the partition function may be expressed as
834: \begin{align}
835:  Z_{mn} (i s) =  \sum_{k= -\infty}^{\infty}\xi ^{k(m-n)}
836:     \frac{2 \sinh (\pi |k| \gamma )\eta (i s) e^{- \pi s k^2 \gamma ^2}
837:   }{\vartheta_1 ( |k|\gamma s|i s) }~,
838:  \label{part}
839: \end{align}
840: which can be regarded as scattering between boundary states 
841: as shown below.
842: 
843: The above expression can be re-derived by path integral method as well. 
844: We use Euclidean worldsheet $(\sigma_1 ,
845: \sigma_2 )$ with $0 \leq \sigma_1 \leq \pi$ and $\sigma_2 \sim
846: \sigma_2 + 2\pi$, where the worldsheet metric and the Laplacian
847: are given by
848: \begin{align}
849:  ds^2 &= d^2 \sigma_1 + t^2 d^2 \sigma_2 ~,
850: &\Delta &= \frac{1}{t^2}(\partial_2^2 + t^2 \partial^2_1) ~.
851: \end{align}
852: We assign Dirichlet condition at the boundaries of worldsheet
853: as
854: \begin{align}
855:  X^{\pm} (\sigma_1=0 , \sigma_2) &=0 ~,
856: &X^{\pm} (\sigma_1= \pi , \sigma_2) &=0 ~.
857:  \label{D1bcpi1}
858: \end{align}
859: Under the identification of discrete boost \eqref{misner},
860: $k$-th twisted sector along $\sigma_2$ obeys
861: \begin{align}
862:  X^{\pm} (\sigma_1 , \sigma_2 + 2 \pi) =
863:   e^{\pm 2 \pi \gamma k} X^{\pm} (\sigma_1 , \sigma_2) ~.
864:  \label{D1bcpi2}
865: \end{align}
866: Conditions \eqref{D1bcpi1} and \eqref{D1bcpi2} altogether are solved by
867: \begin{align}
868:  X^{\pm} (\sigma_1, \sigma_2) =
869: \sum_{m \in {\mathbb Z},n > 0} a^{\pm}_{m,n}
870:   e^{i (m \mp i k\gamma) \sigma_2}\sin (n\sigma_1) ~.
871: \end{align}
872: Assuming the open string has the Chan-Paton
873: indices $(m,n)$ as before, the prescription in \eqref{CP} gives an overall factor $\xi^{k(m-n)}$ to the $k$-th
874: twisted sector.
875: Therefore, the partition function is summarized as%
876: \footnote{See, e.g., \cite{HNP2} for the detailed calculation.}
877: \begin{align}
878: \begin{aligned}
879: Z_{mn} (it) &= \sum_{k = - \infty}^{\infty} \xi^{k(m-n)}  {\rm Det}^{-1} (- \Delta ) \\
880:  & = \sum_{k = - \infty}^{\infty} \xi^{k(m-n)} \left[ \prod_{m,n > 0} (m+itn-ik\gamma)(m-itn-ik\gamma) \right]^{-1} ~,
881: \end{aligned}
882: \end{align}
883: which reproduces \eqref{parto}.
884: 
885: Next, we construct boundary states, which show how closed
886: strings couple to instantons.
887: Since the winding strings have zero size at the big crunch/big
888: bang singularity, the fractional instantons can couple to these 
889: winding strings.
890: Boundary states in the $k$-th twisted sector have to satisfy
891: \begin{align}
892: (\alpha^\pm_{n} - \tilde{\alpha}^\pm_{-n}) 
893: | B_{x^+,x^-}, k \rangle\rangle = 0 
894: \label{inst_cond}
895: \end{align}
896: for all $n \in {\mathbb Z}$, which
897: implies that the instantons can couple to
898: winding strings wrapping $\psi$-cycle for massive
899: strings and $\chi$-cycle for tachyonic strings.
900: It is nice for us since we would like to condense the
901: tachyonic modes of this type.
902: 
903: The boundary state satisfying the condition 
904: $\eqref{inst_cond}$ is given by
905: \begin{align}
906: | B_{x^+,x^-}, k \rangle\rangle 
907:  = \exp \left(-\sum_{n\ge1}\frac{\alpha^+_{-n}\tilde{\alpha}^-_{-n}}{n + i\nu}
908:  - \sum_{n\ge0}\frac{\alpha^-_{-n}\tilde{\alpha}^+_{-n}}{n - i\nu} \right)
909:  |0\rangle ~,
910:  \label{ishibashi}
911: \end{align}
912: whose overlaps are computed as
913: \begin{align}
914: \langle\langle B_{x^+,x^-}, k |e^{-\pi{s}(L_0+\tilde{L}_0-\frac{1}{6})}| B_{x^+,x^-}, k ' \rangle\rangle 
915:           = \delta_{k,k'}
916:     \frac{\eta (is) e^{-\pi s k^2 \gamma ^2}}{\vartheta_1 (s | k| \gamma | is)} ~.
917: \end{align}
918: The boundary state for fractional instanton with label $n$
919: can be given in a linear combination of \eqref{ishibashi}.
920: The coefficients are determined by the fact that
921: the partition function \eqref{part} should be
922: written in terms of boundary states as
923: \begin{align}
924:  Z_{m,n} (i s) =
925:  \langle\langle m |e^{-\pi{s}(L_0+\tilde{L}_0-\frac{1}{6})}| n \rangle\rangle ~.
926: \end{align}
927: The results are
928: \begin{align}
929:   | n \rangle\rangle 
930: &= \sum_{k=-\infty}^{\infty} \xi^{-kn} {\cal N}_k  
931:      | B_{x^+,x^-}, k \rangle\rangle ~,
932:   &\langle \langle n |
933: &= \sum_{k=-\infty}^{\infty} \xi^{kn}  {\cal N}_k 
934:      \langle \langle B_{x^+,x^-}, k | 
935: \end{align}
936: with ${\cal N}_k = \sqrt{ 2 \sinh (  \pi |k| \gamma ) }$.
937: In order to deal with D-instantons in critical bosonic string
938: theory, the sector with $24$ free boson ought to be added.
939: The superstring generalization is demonstrated in appendix B.
940: 
941: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
942: 
943: \subsection{Effective action of open strings on instantons}
944: \label{matrix}
945: 
946: We are going to build low energy effective action of
947: open strings on the fractional instantons presented above.
948: As shown above, there are tachyonic modes in 
949: either bosonic strings or superstrings with opposite
950: GSO projection to odd twisted sectors.
951: For a while we concentrate on the bosonic sector.
952: As a probe, we utilize D$p$-instanton,
953: which is a point-like object in Misner space but
954: $p+1$ dimensional extended object in the extra $d$ directions.%
955: \footnote{The notation we take is $I=\pm,2,3,\cdots,d+2$
956: for total space and $\mu = 2,3, \cdots, p+2$ for parallel 
957: directions to D-instanton. For normal directions, 
958: we use $A=\pm,a$ with $a = p+3 , \cdots , d+2$.}
959: As in the previous case, we start from the bulk instanton.
960: In the covering space we have to introduce infinitely many
961: image D-instantons, and hence the low energy effective theory
962: has $\infty \times \infty$ matrices
963: $X^{A}_{i,j}$, $A^{\mu}_{i,j}$, where $i,j$
964: are Chan-Paton indices.
965: 
966: 
967: When it comes to fractional instanton case,
968: we rather change the basis of Chan-Paton indices \eqref{basis} as
969: \begin{align}
970: X^A_{m,n} &= 
971:  \sum_{i,j = - \infty}^{\infty} \xi^{im} X^A_{i,j} \xi^{-jn} ~,
972: &A^{\mu}_{m,n} &= 
973:  \sum_{i,j = - \infty}^{\infty} \xi^{im} A^{\mu}_{i,j} \xi^{-jn} ~.
974: \end{align}
975: The orbifold action acts on these matrices as
976: \begin{align}
977: g : X^{\pm}_{i,j} &\to e^{\pm 2 \pi \gamma} X^{\pm}_{i-1,j-1} ~,
978: &g : A^{\mu}_{i,j} &\to A^{\mu}_{i-1,j-1} ~,
979: &g : X^{a}_{i,j} &\to X^{a}_{i-1,j-1} ~,
980: \end{align}
981: or in the new basis as 
982: \begin{align}
983: g : X^{\pm}_{m,n} &\to \xi^{m - n} e^{\pm 2 \pi \gamma} X^{\pm}_{m,n} ~,
984: &g : A^{\mu}_{m,n} &\to \xi^{m - n} A^{\mu}_{m,n} ~,
985: &g : X^{a}_{m,n} &\to \xi^{m - n} X^{a}_{m,n} ~.
986: \end{align}
987: {}From now on we assign $\xi = e^{2 \pi \gamma}$, then
988: only $X^{\pm}_{n , n \pm 1}$, $A^{\mu}_{n,n}$, $X^a_{n,n}$
989: survive the orbifold projection.
990: This is a similar situation to Euclidean orbifold case
991: \cite{DM} even though the indices do not obey periodicity condition.
992: If we use generic $\xi$, then all $X^{\pm}_{m,n}$ are projected
993: out.\footnote{If we choose $\xi = e^{2 \pi \gamma /N}$
994: with $N \in {\mathbb Z}$, then $X^{\pm}_{m,m+N}$ would survive.
995: Since the space-time fermions $S^{\pm}_{m,n}$ in our superstring 
996: case transform as 
997: $g:S^{\pm}_{m,n} \to - \xi^{m-n} e^{\pm \pi \gamma} S^{\pm}_{m,n}$,
998: we may choose $\xi = - e^{\pi \gamma}$ to have 
999: non-trivial components $S^{\pm}_{m,m \pm 1}$ 
1000: along with $X^{\pm}_{m,m\pm 2}$.} 
1001: 
1002: 
1003: In general, the effective Lagrangian for the matrices on 
1004: D$p$-instantons is given by
1005: \begin{align}
1006:  {\cal L} = -  \frac14 {\rm Tr} \, 
1007:  \left( F_{\mu \nu} F^{\mu \nu} + 2 D_{\mu} X_A D^{\mu} X^A
1008:     - [ X_A , X_B ] ^2
1009:  \right) ~,
1010:  \label{matrixaction}
1011: \end{align}
1012: where the trace runs over Chan-Paton indices.
1013: For simplicity, we neglect $g$ and $2 \pi \alpha '$ in the Lagrangian.
1014: Now we choose a configuration with every fractional
1015: instantons labeled by $m \in {\mathbb Z}$.
1016: Here we should remark that the Lagrangians with the bases $(i,j)$
1017: and $(m,n)$ might not be the same because the transformation of
1018: basis is not unitary.
1019: In fact, the both Lagrangians would not coincide unless
1020: $\sum _k {\rm Tr} \xi^{kn} = 0$, which is true for 
1021: ${\mathbb C}/{\mathbb Z}_N$ case, for instance.%
1022: \footnote{It might be possible that the bulk instanton and
1023: the sum of every fraction instantons are regarded as the same object
1024: if we choose a proper regularization, which we do not know yet.}
1025: 
1026: 
1027: 
1028: Applying the orbifold projection, 
1029: the Lagrangian involving the scalars $X^{\pm}$
1030: is given by
1031: \begin{align}
1032: \begin{aligned}
1033:  {\cal L}_{\pm} = &
1034:    \sum_{n}  \Big[ ( \partial_{\mu} - B_{\mu,n} ) 
1035:      X^+_{n,n+1} ( \partial_{\mu} + B_{\mu,n} )   X^-_{n+1,n}  \\
1036:   & \qquad -   (X^a_{n,n} - X^a_{n+1,n+1})^2  
1037:     | X^+_{n,n+1} |^2 + \frac12 
1038:       ( | X^+_{n,n+1} |^2 - |X^+_{n-1,n}|^2)^2 \Big] ~,
1039: \end{aligned}
1040: \end{align}
1041: where we have defined 
1042: $B_{\mu,n} = i (A_{\mu,n,n} -  A_{\mu,n+1,n+1})$ and 
1043: $|X^+_{n,n+1}|^2 = X^{+}_{n,n+1} X^-_{n+1,n}$.
1044: Due to the orbifold projection, the gauge symmetry left is
1045: \begin{align}
1046:  B_{\mu,n} &\to B_{\mu,n} + \partial_{\mu} \Lambda_n (x^{\mu}) ~,
1047: &X^{+}_{n,n+1} &\to e^{\Lambda_n (x^{\mu})} X^{+}_{n,n+1} ~,
1048: &X^{-}_{n+1,n} &\to e^{-\Lambda_n (x^{\mu})} X^{-}_{n+1,n} ~.
1049: \label{gauge}
1050: \end{align}
1051: We can read off the classical moduli space from 
1052: the saddle point of the potential term as%
1053: \footnote{As mentioned before, we focus on the big bang region
1054: ($X^+ , X^- > 0$) of Misner space.}
1055: \begin{align}
1056: X^{\pm}_{n,n\pm 1} &= 0 &{\rm or }& &X^a_{n,n} &= x^a,
1057:   \quad | X^+_{n,n+1} |^2 = \frac{t^2}{2} ~.
1058: \end{align}
1059: The former corresponds to Coulomb branch, where the fractional
1060: instantons are localized at the big bang singularity as expected.
1061: The latter corresponds to Higgs branch, where the center of
1062: the sum of fractional instantons can be placed out of the fixed point.
1063: 
1064: 
1065: The situation bears resemblance to the case of ${\mathbb C}/{\mathbb Z}_N$.
1066: There the sum of $N$ different fractional branes can be moved 
1067: away from the fixed point, since the bulk brane is just the sum of them.
1068: Likewise, we can probe the geometry away from the singularity
1069: by using the sum of every fractional instantons, even though
1070: it might differ from the bulk instanton.
1071: 
1072: 
1073: This may be understood as follows.
1074: In the boundary state formalism, we can insert a operator into the
1075: boundary state
1076: \begin{align}
1077:   {\rm Tr} {\rm P} \exp \left( - i \int d \sigma M_A \partial_{\tau} X^{A} (\sigma ) \right) ~,
1078: \label{wilson}
1079: \end{align}
1080: where P represents path ordering.
1081: This operator is T-dual to the Wilson line and generates the shift
1082: of the position of corresponding brane.
1083: To preserve the conformal invariance, it is necessary to impose some conditions to $M_A$ .
1084: In the first order of $\alpha '$, the condition that
1085: the beta function vanishes is equal to the equations of
1086: motion to the effective Lagrangian;
1087: $[ M_A , M_B ] = 0$ ($|X^{\pm}_{n,n+1}|^2 = t^2 /2$).
1088: To conclude, even the composite fractional instantons 
1089: cannot be moved by themselves, the center of the sum of them can be shifted
1090: by the insertion of the shift operator \eqref{wilson}.
1091: 
1092: 
1093: Let us turn to examine moduli space on the Higgs branch.
1094: On this branch, we can set 
1095: $X^{+}_{n,n+1} = \frac{1}{\sqrt{2}}t e^{ \psi_n}$ and
1096: $X^{-}_{n+1,n} = \frac{1}{\sqrt{2}}t e^{- \psi_n}$.
1097: Making use of the gauge symmetry \eqref{gauge}, we may fix
1098: $\psi_n = \psi$.
1099: This gauge choice leaves unfixed ${\mathbb Z}$ symmetry
1100: as in \cite{APS}, which is generated by
1101: $\exp ( - 2 \pi \gamma \sum_n n Q_n )$.
1102: This means that we have to identify 
1103: $X^{\pm} \sim e^{\pm 2 \pi \gamma} X^{\pm}$, namely
1104: the coordinate $\psi$ has periodicity
1105: $\psi \sim \psi + 2 \pi \gamma$.
1106: In this way, we can identify that the moduli space is 
1107: Misner space as is expected.
1108: 
1109: 
1110: The metric of the moduli space is obtained by means of ``kinetic'' term of the Lagrangian. It can be computed as
1111: \begin{align}
1112:  - \sum_{n}  ( \partial_{\mu} - B_{\mu,n} ) 
1113:      X^+_{n,n+1} ( \partial_{\mu} + B_{\mu,n} ) 
1114:      X^-_{n+1,n}   = 
1115:   \sum_i \frac{1}{2} \left[ - (\partial_{\mu} t)^2 + 
1116:  t^2 (\partial_{\mu} \psi_n - B_{\mu , n} )^2 \right] ~.
1117: \end{align}
1118: Decoupling diagonal gauge symmetry forces us to assign $\sum_n B_{\mu , n} = 0$.
1119: This restriction is incorporated into the Lagrangian by 
1120: introducing the Lagrange multiplier 
1121: $\lambda_{\mu}$. Solving the equation of motion for 
1122: $B_{\mu , n}$ and the constraint $\sum_n B_{\mu , n} = 0$, we obtain
1123: \begin{align}
1124:  B_{\mu , n} &= \partial_{\mu} ( \psi_n - \psi ) ~,
1125:  &\sum_n \psi_n = \sum_n (\psi + 2 \pi \gamma n) ~.
1126: \end{align}
1127: In this definition of $\psi$, we have desired periodicity 
1128: $\psi \sim \psi + 2 \pi \gamma$. Then, the kinetic term
1129: becomes
1130: \begin{align}
1131:  \frac{1}{2} \left[ - (\partial_{\mu} t)^2 + 
1132:  t^2 (\partial_{\mu} \psi )^2 \right] ~,
1133: \end{align}
1134: which reproduces the metric in the cosmological region of 
1135: Misner space. 
1136: 
1137: \subsection{Closed string tachyon condensation}
1138: 
1139: Taking advantage of the effective theory constructed above,
1140: we are capable of discussing the condensation of closed string
1141: tachyons in the twisted sectors.
1142: Since we are interested in the region near the singularity,
1143: it is enough to think of the deformation to 
1144: leading terms of $X^{\pm}$ as
1145: \begin{align}
1146:  \Delta V = \sum_n m_n^2 |X^+_{n,n+1}|^2  ~.
1147: \end{align}
1148: This term should originate from the three point function
1149: like $\langle \sigma_k \psi^+ \psi^- \rangle$,
1150: where $\sigma_k$ stands for the $k$-th twist vertex operator
1151: of closed string \cite{DM}.
1152: Put it differently, bringing vev to the tachyonic mode causes
1153: the mass deformation.
1154: The three point function would not vanish since the sum of fractional
1155: branes couples to the twisted closed strings as shown before.
1156: It is here of practical advantage to probe the process by fractional instantons,
1157: since this property would be far from obvious in the case of bulk instanton.
1158: 
1159: 
1160: For the purpose of the following argument, it helps to re-phrase the mass 
1161: terms as
1162: \begin{align}
1163:  \Delta V = - \sum_n \lambda_n (|X^+_{n,n+1}|^2 - |X^+_{n-1,n}|^2) ~.
1164:  \label{pb}
1165: \end{align}
1166: Then, the saddle point of the potential is shifted as
1167: \begin{align}
1168:  |X^+_{n,n+1}|^2 - |X^+_{n-1,n}|^2 = \lambda_n ~.
1169:  \label{FI}
1170: \end{align}
1171: Hence, $| X^+_{n,n+1}|^2 = t^2_n/2$ depends on $n$ 
1172: after the tachyon condenses.
1173: 
1174: 
1175: 
1176: Let us first examine the simplest case, say,
1177: only $\lambda_1 = \rho^2/2$ is non-zero.
1178: Then, we can just set
1179: \begin{align}
1180:  |X^+_{n,n+1}|^2 & = \frac{t^2}{2}  \quad {\rm for} ~ n \leq 0 ~,
1181:  &|X^+_{n,n+1}|^2 &=  \frac{t^2}{2} + \frac{\rho^2}{2} \quad  {\rm for} ~ n > 0 ~.
1182: \end{align}
1183: The above equations imply that the ungauged 
1184: ${\mathbb Z}$ symmetry is reduced to ${\mathbb Z}/2$ symmetry.
1185: Following the previous analysis, we obtain the metric of 
1186: moduli space as
1187: \begin{align}
1188:  &\frac{1}{2} \left[ - A( t ) 
1189:   (\partial_\mu t )^2 + \frac{t^2}{A( t )} (\partial_{\mu} \psi)^2
1190:   \right]~, 
1191:   &A( t ) &=  \frac{1}{2}\left( 1 + \frac{t^2}{t^2 + \rho^2} \right)~.
1192: \end{align}
1193: Notice that the function $A( t )$ connects 1 at $|t|=\infty $ 
1194: and $\frac{1}{2}$ at $|t|=0$. 
1195: In conclusion, Misner space with periodicity $\psi \sim \psi + 2 \pi \gamma$
1196: is now replaced by another Misner space with parameter 
1197: $\psi \sim \psi + 4 \pi \gamma$ near the big bang singularity.
1198: In this way, we can understand that the singularity gets milder by
1199: the tachyon condensation,\footnote{Precisely speaking,
1200: there is still curvature singularity even in Misner space with
1201: larger space cycle. The singularity is resolved if the final
1202: space-time is a part of Minkowski space-time as in the next example.}
1203: or the time period is shortened.%
1204: \footnote{Let us compare two Misner spaces with the periodicity 
1205: $2 \pi \gamma$ and $4 \pi \gamma$ as this example. 
1206: Since the space radius becomes $\gamma L$ when $t=L$ and $t=L/2$, 
1207: respectively,
1208: we interpret this fact as the effective time period is shortened by
1209: the tachyon condensation. This interpretation is consistent with
1210: that in \cite{MS}, where it was claimed that the time-like Liouville 
1211: potential repels the wave function near the big bang singularity.
1212: }
1213: As shown in \cite{APS}, the tip of the cone of
1214: ${\mathbb C}/{\mathbb Z}_N$ model is smoothed out by tachyon
1215: condensation, and our case can be regarded as an analogous situation.
1216: Repeating similar deformations, Misner space 
1217: with periodicity $\psi \sim \psi + 2 \pi \gamma$ flows to
1218: Misner space
1219: with periodicity $\psi \sim \psi + 2 \pi \gamma c$ with $c > 1$.
1220: 
1221: 
1222: We are in a position to deal with more generic $\lambda_n$.
1223: We can set $|X^+_{0,1}|^2= t^2/2$ to be the smallest one among all
1224: by re-labeling $n$. Defining $\rho_n$ by 
1225: $\rho_{n}^2 - \rho^2_{n-1} = 2 \lambda_n$ with $\rho_0 = 0$,
1226: the solution to \eqref{FI} is given by 
1227: $|X^{\pm}_{n,n+1}|^2 = t_n^2 /2 = (t^2 + \rho_n^2)/2 $.
1228: Due to the generic deformation, the ${\mathbb Z}$ symmetry is completely
1229: broken. Using $N = \sum_n 1$ (which has been neglected elsewhere), 
1230: the kinetic term turns out to be 
1231: \begin{align}
1232:  &\frac{N}{2} \left[ - A( t ) 
1233:   (\partial_\mu t )^2 + \frac{t^2}{A( t )} (\partial_{\mu} \psi)^2
1234:   \right]~, 
1235:   &A( t ) &=  \frac{1}{N} \sum _n \left(\frac{t^2}{t^2 + \rho_n^2} \right)~.
1236: \end{align}
1237: The function $A( t )$ interpolates 1 at $|t|=\infty $ 
1238: and $0$ at $|t|=0$.
1239: Thus the region near the origin of Misner space with 
1240: $2 \pi \gamma$ as the periodicity of $\psi$-cycle is replaced by
1241: Misner space with $\infty$ $\psi$-cycle, namely, 
1242: the future patch of Minkowski space in the Rindler coordinates.
1243: Now that the time-coordinate is shortened by tachyon condensation,
1244: the Minkowski region is not connected to the other regions in
1245: a simple way.
1246: 
1247: 
1248: 
1249: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1250: \section{Conclusion} 
1251: \label{conclusion}
1252: 
1253: We have explored the condensation of 
1254: twisted closed string tachyon in Misner space
1255: accompanied by D-instanton probe.
1256: We spotlight on the big bang region with $t > 0$, that is, the problem involved is whether 
1257: tachyon condensation resolves big bang singularity.
1258: In the bosonic string or superstring theories with
1259: opposite periodic condition of space-time fermion, 
1260: there appears tachyonic modes in the twisted
1261: sectors of closed string.
1262: We have shown that a type of closed string
1263: tachyon is localized in the whisker regions
1264: with the help of classical string trajectories
1265: and Virasoro constraints.
1266: The condensation of the twisted tachyon
1267: affects the cosmological region through the
1268: big bang singularity, which bears a strong resemblance to the localized
1269: tachyon condensation in Euclidean orbifolds \cite{APS}.
1270: 
1271: 
1272: In order to see how the small vev of the tachyon changes
1273: the geometry, it is useful to utilize D-instanton probe.
1274: There are ``fractional'' instantons in Misner space, and it is
1275: the sum of them that provides us with a powerful tool to gain further insight.
1276: Two main properties are responsible for this;
1277: (i) D-instanton can couple to closed string tachyon in the twisted sectors, 
1278: and (ii) it is able to be moved away from the big bang singularity.
1279: {}Resorting to the effective action on D-instanton,
1280: it is possible to read off the metric of the background 
1281: both before and after the tachyon condensation.
1282: 
1283: 
1284: After tachyon condenses, it has been shown that Misner space with larger space cycle replaces
1285: Misner space with smaller one near the big bang singularity.
1286: In generic tachyon condensation, Misner space is 
1287: replaced by the future patch of Minkowski space in Rindler coordinates.
1288: The tachyon condensation shortens the effective period of time
1289: \cite{MS}, and hence the beginning of the space-time is not directly 
1290: connected with the other regions.
1291: 
1292: Notice that D-instanton probe stays
1293: reliable merely in the region less than string length \cite{DKPS}.
1294: Therefore, we cannot examine how the tachyon condensation changes
1295: the geometry beyond this region by D-instanton probe.
1296: Far away from the big bang singularity, any geometry change should develop in the gravity regime \cite{APS}.
1297: It may be possible to study this regime by following RG flow or referring to time dynamics.
1298: 
1299: 
1300: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1301: \subsection*{Acknowledgement}
1302: 
1303: We would like to thank K.~Hosomichi, K.~Ideguchi, Y.~Matsuo, 
1304: S.~Nagaoka, F.~Yagi and A.~Yamaguchi for useful discussions. 
1305: We also thank the organizers of YITP workshop
1306: ``String Theory and Quantum Field Theory'',
1307: where useful discussions were made.
1308: 
1309: 
1310: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1311: 
1312: \appendix
1313: 
1314: \section{Theta and eta functions}
1315: \label{formulae}
1316: 
1317: We have used the definition of theta functions as
1318: \begin{align}
1319: \begin{aligned}
1320:  \vartheta_1 (\nu|\tau) &= 2 q^{\frac18} \sin \pi \nu
1321:   \prod_{m=1}^{\infty} (1-q^m)(1-zq^m)(1-z^{-1}q^m) ~,
1322:  \\
1323:  \vartheta_2  (\nu|\tau) &= 2 q^{\frac18} \cos \pi \nu
1324:   \prod_{m=1}^{\infty} (1-q^m)(1+zq^m)(1+z^{-1}q^m) ~,
1325:  \\
1326:  \vartheta_3  (\nu|\tau) &=
1327:   \prod_{m=1}^{\infty} (1-q^m)(1+zq^{m-\frac12})(1+z^{-1}q^{m-\frac12}) ~,
1328:  \\
1329:  \vartheta_4  (\nu|\tau) &=
1330:   \prod_{m=1}^{\infty} (1-q^m)(1-zq^{m-\frac12})(1-z^{-1}q^{m-\frac12}) ~,
1331: \end{aligned}
1332: \end{align}
1333: where we have set $q=\exp ( 2 \pi i \tau )$ and $z = \exp (2 \pi i \nu )$.
1334: Their modular transformations are
1335: \begin{align}
1336: \begin{aligned}
1337:  \vartheta_1 \left(\frac{\nu}{\tau} \left| - \frac{1}{\tau} \right.\right)
1338:   &= - i ( - i \tau )^{\frac{1}{2}}
1339:     e^{ \frac{\pi i \nu ^2}{\tau} }
1340:     \vartheta_1 ( \nu | \tau ) ~,
1341:  &\vartheta_2 \left(\frac{\nu}{\tau} \left| - \frac{1}{\tau} \right.\right)
1342:   &= ( - i \tau )^{\frac{1}{2}}
1343:     e^ { \frac{\pi i \nu ^2}{\tau} }
1344:     \vartheta_4 ( \nu | \tau ) ~, \\
1345:  \vartheta_3 \left(\frac{\nu}{\tau} \left| - \frac{1}{\tau} \right.\right)
1346:   &= ( - i \tau )^{\frac{1}{2}}
1347:     e^{ \frac{\pi i \nu ^2}{\tau} }
1348:     \vartheta_3 ( \nu | \tau ) ~,
1349:  &\vartheta_4 \left(\frac{\nu}{\tau} \left| - \frac{1}{\tau} \right.\right)
1350:   &= ( - i \tau )^{\frac{1}{2}}
1351:     e^{ \frac{\pi i \nu ^2}{\tau} }
1352:     \vartheta_2 ( \nu | \tau ) ~.
1353: \end{aligned}
1354: \end{align}
1355: We have also used Dedekind eta function defined as
1356: \begin{align}
1357:  \eta (\tau) &= q^{\frac{1}{24}} \prod_{m=1}^{\infty} (1 - q^m) ~,
1358:  &\eta \left( - \frac{1}{\tau} \right)
1359:  &= ( - i \tau )^{\frac12 } \eta ( \tau ) ~.
1360: \end{align}
1361: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1362: \section{D-instantons in superstring theory}
1363: 
1364: It is also of illustrative interest to construct D-instanton boundary state in superstring cases, 
1365: though the analysis having been made concerns mainly the bosonic part.
1366: In the open string channel the Dirichlet boundary condition is specified as
1367: \begin{align}
1368: \psi^{\pm}  &= - \tilde \psi^{\pm} \quad \text{ at } \sigma = 0 ~,
1369: &\psi^{\pm} &= (-1)^A \tilde \psi^{\pm} \quad \text{ at } \sigma = \pi ~,
1370: \end{align}
1371: where $A=0$ for NS-sector and $A=1$ for R-sector. Through doubling trick, mode expansion is simply 
1372: \begin{align}
1373:  \psi^{\pm} &= \sum_{r}
1374:   b^{\pm}_{-r} e^{-ir(\tau + \sigma)} 
1375: \end{align}
1376: with $r \in {\mathbb Z} + (A+1)/2$.
1377: The oscillators satisfy the anti-commutation relations 
1378: $\{ b^{\pm}_r , b^{\mp}_s \} = - \delta_{r+s}$. 
1379: The orbifold action is defined as
1380: $e^{2 \pi i \gamma \hat J} \psi^{\pm} = e^{\pm 2 \pi \gamma} \psi^{\pm}$, where
1381: \begin{align}
1382:  i J &= - \sum_{r \geq \frac12} b^+_{-r} b^-_r 
1383:        +  \sum_{r \geq \frac12} b^-_{-r} b^+_r
1384:  \end{align}
1385: for NS sector and 
1386: \begin{align}
1387:  i J &= - \sum_{n \geq 1} b^+_{-n} b^-_n 
1388:        +  \sum_{n \geq 0} b^-_{-n} b^+_n
1389:  \end{align}
1390: for R sector.
1391: Equipped with these it is now straightforward to compute the open string one-loop amplitude.%
1392: \footnote{As before, an overall factor $1/N$ with $N=\sum_i 1$ is neglected and also $k=0$ sector is.}
1393: While embedded into the 10d spacetime, including both ghost part 
1394: and GSO projection gives%
1395: \footnote{The factor $(-1)^k$ inside the trace over R-sector comes
1396: from the fact that the space-time fermion has identification
1397: such as $S^{\pm} \sim g^k \cdot S^{\pm}
1398:  = (-1)^k e^{\pm k \pi \gamma} S^{\pm}$ as mentioned before.}
1399: \begin{align}
1400: Z_{mn}(it) &= {\rm Tr}_{{\cal H}_{mn}^{\rm NS}} \sum_k g^k {\textstyle \frac12} (1 + (-1)^F)
1401:   e^{-2\pi t L_0 } - {\rm Tr}_{{\cal H}_{mn}^{\rm R}} \sum_k (-g)^k {\textstyle \frac12} (1 \pm (-1)^F)
1402:   e^{-2\pi t L_0} \nonumber \\
1403: &= \frac{i}{2}\sum^{\infty}_{k = -\infty}\xi^{k({m-n})} 
1404:   2 \sinh (\pi |k| \gamma ) \label{supZ} \\
1405: &\times\frac{\vartheta_4 (i | k |\gamma|it) \vartheta_4 (0|it)^3
1406:              - \vartheta_3(i |k |\gamma|it) \vartheta_3(0|it)^3
1407:              - (-1)^k\vartheta_2 (i |k |\gamma|it) \vartheta_2 ( 0 |it)^3}
1408:              {(8\pi^2 t \alpha')^{\frac{p+1}{2}} 
1409:               \vartheta_1(i |k |\gamma|it) \eta(it)^9} ~, 
1410:               \nonumber
1411: \end{align}
1412: where Neumann boundary condition for
1413: $p+1$ coordinates and Dirichlet boundary condition for $7-p$
1414: coordinates are assumed.
1415: Under the modular transformation $t\rightarrow s=1/t$,
1416: the partition function is changed into
1417: \begin{align}
1418: \begin{aligned}
1419:  &Z_{mn}(is) = \frac{1}{2}\sum^{\infty}_{k = -\infty}\xi^{k({m-n})}2 \sinh (\pi |k| \gamma)\\
1420:  & \quad \times\frac{\vartheta_3 ( | k | \gamma s|is) \vartheta_3 (0|is)^3
1421:              - (-1)^k\vartheta_4 ( | k | \gamma s|is) \vartheta_4 (0|is)^3
1422:              - \vartheta_2 ( | k | \gamma s|is) \vartheta_2 ( 0 |is)^3}
1423:              {(8\pi^2 \alpha')^{\frac{p+1}{2}}s^{\frac{7-p}{2}}
1424:                \vartheta_1( | k | \gamma s|is) \eta(is)^9} ~,
1425:                \label{superD1}
1426: \end{aligned}
1427: \end{align}
1428: which should be reproduced by the overlap between D-instanton boundary states.
1429: 
1430: Fermionic contribution to the boundary states in $k$-th twisted 
1431: sector of Misner part 
1432: \begin{align}
1433: \begin{aligned}
1434:  ( b^{\pm}_{-r} - i \eta \tilde b^{\pm}_r ) 
1435: | B_{\psi^{\pm}}^f, k, \eta \rangle\rangle_\text P = 0 ~, 
1436: &&\eta = \pm1
1437: \end{aligned}
1438: \end{align}
1439: with $P$ referred to NSNS or RR is solved by
1440: \begin{align}
1441: \begin{aligned}
1442: | B_{\psi^{\pm}}^f, k, \eta \rangle\rangle_{\text{NSNS}}
1443:  &= e^{-i \eta \sum_{n \geq 1} \left(b^+_{-n + 1/2} \tilde b^-_{- n + 1/2}
1444:       + b^-_{-n + 1/2} \tilde b^+_{- n + 1/2} \right) } | 0 \rangle  ~,
1445: \\
1446: | B_{\psi^{\pm}}^f, k, \eta \rangle\rangle_{\text {RR}}
1447:  &= e^{-i \eta \left(\sum_{n \geq 1} b^+_{-n} \tilde b^-_{-n}
1448:       + \sum_{n \geq 0} b^-_{-n} \tilde b^+_{- n} \right) } | 0 \rangle ~.
1449: \end{aligned}
1450: \end{align}
1451: Gathering all components to form the GSO
1452: invariant boundary state, it is thus of the form
1453: \begin{align}
1454: \begin{aligned}
1455:  | B, k \rangle\rangle =& 
1456:   \frac12 \left( | B, k, + \rangle \rangle_{\text {NSNS}}
1457:     - (-1)^k | B, k, - \rangle \rangle_{\text {NSNS}} \right)  \\
1458:     &+ \frac12 \left( | B, k, + \rangle \rangle_{\text {RR}}
1459:     + (-1)^k | B, k, - \rangle \rangle_{\text {RR}} \right) 
1460: \end{aligned}
1461: \end{align}
1462: with
1463: \begin{align}
1464: \begin{aligned}
1465:  | B, k, \eta \rangle\rangle_P = {\cal N}_k | B_{\pm}, k, \eta \rangle\rangle_P \otimes | B_{\text {flat}}, \eta \rangle\rangle_P \otimes
1466:  | B_{\text {gh}}, \eta \rangle\rangle_P ~.
1467: \label{Coeff}
1468:  \end{aligned}
1469: \end{align}
1470: See \cite{HNP2} for the definition of extra direction part.
1471: The overlap of the boundary states is demanded to be equal to 
1472: the open string one-loop amplitude such that
1473: \begin{align}
1474: Z_{mn}  = 
1475:  \frac{\alpha ' \pi}{2} \int_0^{\infty} ds
1476:  \langle \langle m|e^{-\pi{s}(L_0+\tilde{L}_0)}|n\rangle \rangle ~,
1477: \end{align}
1478: where 
1479: \begin{align}
1480: | m \rangle \rangle 
1481:  = \sum_{k=-\infty}^{\infty} B_m^k | B, k \rangle\rangle
1482: \end{align}
1483: is defined.
1484: {}From the above requirement, we obtain
1485: \begin{align}
1486:  B_m^k &= \xi^{- k m} ~,
1487:  &{\cal N}_k &=
1488:  \sqrt{\frac{2 \sinh (\pi |k| \gamma)}{2 \pi ^2 \alpha '}} {\cal N}_p~,
1489: &{\cal N}_p &= \frac{\sqrt{\pi}}{2} (2 \pi \sqrt{\alpha '})^{3 - p} ~,
1490: \end{align}
1491: where ${\cal N}_p$ is the coefficient of the boundary states
1492: for usual D$p$-brane. For bra states, we should use $B_m^{- k}$
1493: instead of $B_m^k$.
1494: 
1495: 
1496: 
1497: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1498: 
1499: \baselineskip=14pt
1500: %\small
1501: \begin{thebibliography}{99}
1502: 
1503: \bibitem{KOSST}
1504:   J.~Khoury, B.~A.~Ovrut, N.~Seiberg, P.~J.~Steinhardt and N.~Turok,
1505:    ``From big crunch to big bang,''
1506:   Phys.\ Rev.\ D {\bf 65}, 086007 (2002)
1507:   [arXiv:hep-th/0108187].
1508:   %%CITATION = HEP-TH 0108187;%%
1509: 
1510: \bibitem{BHKN}
1511:   V.~Balasubramanian, S.~F.~Hassan, E.~Keski-Vakkuri and A.~Naqvi,
1512:   ``A space-time orbifold: A toy model for a cosmological singularity,''
1513:   Phys.\ Rev.\ D {\bf 67}, 026003 (2003)
1514:   [arXiv:hep-th/0202187].
1515:   %%CITATION = HEP-TH 0202187;%%
1516: 
1517: \bibitem{CC}
1518:   L.~Cornalba and M.~S.~Costa,
1519:   ``A new cosmological scenario in string theory,''
1520:   Phys.\ Rev.\ D {\bf 66}, 066001 (2002)
1521:   [arXiv:hep-th/0203031].
1522:   %%CITATION = HEP-TH 0203031;%%
1523: 
1524: \bibitem{Nekrasov}
1525:   N.~A.~Nekrasov,
1526:   ``Milne universe, tachyons, and quantum group,''
1527:   Surveys High Energ.\ Phys.\  {\bf 17}, 115 (2002)
1528:   [arXiv:hep-th/0203112].
1529:   %%CITATION = HEP-TH 0203112;%%
1530: 
1531: \bibitem{Simon}
1532:   J.~Simon, ``The geometry of null rotation identifications,'' JHEP
1533:   {\bf 0206}, 001 (2002) [arXiv:hep-th/0203201].
1534:   %%CITATION = HEP-TH 0203201;%%
1535: 
1536: \bibitem{LMS1}
1537:   H.~Liu, G.~W.~Moore and N.~Seiberg,
1538:   ``Strings in a time-dependent orbifold,''
1539:   JHEP {\bf 0206}, 045 (2002)
1540:   [arXiv:hep-th/0204168].
1541:   %%CITATION = HEP-TH 0204168;%%
1542: 
1543: \bibitem{CCK} L. Cornalba, M.S. Costa, and C. Kounnas,
1544:   ``A resolution of the cosmological singularity with
1545:   orientifolds'', Nucl. \ Phys. \ B{\bf 637} 378 (2002) arXiv:
1546:   hep-th/0204261.
1547:   %%CITATION = HEP-TH 0204261%%
1548: 
1549: \bibitem{Lawrence}
1550:   A.~Lawrence, ``On the instability of 3D null singularities,'' JHEP
1551:   {\bf 0211}, 019 (2002) [arXiv:hep-th/0205288].
1552:   %%CITATION = HEP-TH 0205288;%%
1553: 
1554: \bibitem{LMS2}
1555:   H.~Liu, G.~W.~Moore and N.~Seiberg,
1556:    ``Strings in time-dependent orbifolds,''
1557:   JHEP {\bf 0210}, 031 (2002)
1558:   [arXiv:hep-th/0206182].
1559:   %%CITATION = HEP-TH 0206182;%%
1560: 
1561: \bibitem{FM}
1562:   M.~Fabinger and J.~McGreevy, ``On smooth time-dependent orbifolds
1563:   and null singularities,'' JHEP {\bf 0306}, 042 (2003)
1564:   [arXiv:hep-th/0206196].
1565:   %%CITATION = HEP-TH 0206196;%%
1566: 
1567: \bibitem{HP}
1568:   G.~T.~Horowitz and J.~Polchinski, ``Instability of space-like and
1569:   null orbifold singularities,'' Phys.\ Rev.\ D {\bf 66}, 103512
1570:   (2002) [arXiv:hep-th/0206228].
1571:   %%CITATION = HEP-TH 0206228;%%
1572:   
1573: \bibitem{NW}
1574:   C.~R.~Nappi and E.~Witten,
1575:    ``A closed, expanding universe in string theory,''
1576:   Phys.\ Lett.\ B {\bf 293}, 309 (1992)
1577:   [arXiv:hep-th/9206078].
1578:   %%CITATION = HEP-TH 9206078;%%
1579: 
1580: \bibitem{EGKR}
1581:   S.~Elitzur, A.~Giveon, D.~Kutasov and E.~Rabinovici,
1582:    ``From big bang to big crunch and beyond,''
1583:   JHEP {\bf 0206}, 017 (2002)
1584:   [arXiv:hep-th/0204189].
1585:   %%CITATION = HEP-TH 0204189;%%
1586: 
1587: \bibitem{CKR}
1588:   B.~Craps, D.~Kutasov and G.~Rajesh,
1589:    ``String propagation in the presence of cosmological singularities,''
1590:   JHEP {\bf 0206}, 053 (2002)
1591:   [arXiv:hep-th/0205101].
1592:   %%CITATION = HEP-TH 0205101;%%
1593: 
1594: \bibitem{HT}
1595:   Y.~Hikida and T.~Takayanagi,
1596:    ``On solvable time-dependent model and rolling closed string tachyon,''
1597:   Phys.\ Rev.\ D {\bf 70}, 126013 (2004)
1598:   [arXiv:hep-th/0408124].
1599:   %%CITATION = HEP-TH 0408124;%%
1600: 
1601: \bibitem{TT}
1602:   N.~Toumbas and J.~Troost,
1603:    ``A time-dependent brane in a cosmological background,''
1604:   JHEP {\bf 0411}, 032 (2004)
1605:   [arXiv:hep-th/0410007].
1606:   %%CITATION = HEP-TH 0410007;%%
1607: 
1608: \bibitem{Misner}
1609:   C.~W.~Misner,
1610:   ``Taub-NUT space as a counterexample to almost
1611:   anything,''
1612:   Relativity Theory and Astrophysics I: Relativity and
1613:   Cosmology, ed. J.~Ehlers, Lectures in Applied Mathematics,
1614:   Volume 8, American Mathematical Society, 160 (1967).
1615: 
1616: \bibitem{BCKR}
1617:   M.~Berkooz, B.~Craps, D.~Kutasov and G.~Rajesh,
1618:    ``Comments on cosmological singularities in string theory,''
1619:   JHEP {\bf 0303}, 031 (2003)
1620:   [arXiv:hep-th/0212215].
1621:   %%CITATION = HEP-TH 0212215;%%
1622:   
1623: \bibitem{Pioline1}
1624:   M.~Berkooz and B.~Pioline, ``Strings in an electric field, and the
1625:   Milne universe,'' JCAP {\bf 0311}, 007 (2003)
1626:   [arXiv:hep-th/0307280].
1627:   %%CITATION = HEP-TH 0307280;%%
1628: 
1629: \bibitem{Pioline2}
1630:   M.~Berkooz, B.~Pioline and M.~Rozali, ``Closed strings in Misner
1631:   space: Cosmological production of winding strings,''
1632:   JCAP {\bf 0408}, 004 (2004)
1633:   [arXiv:hep-th/0405126].
1634:   %%CITATION = HEP-TH 0405126;%%
1635: 
1636: \bibitem{Pioline3}
1637:   M.~Berkooz, B.~Durin, B.~Pioline and D.~Reichmann,
1638:   ``Closed strings in Misner space: Stringy fuzziness with a twist,''
1639:   JCAP {\bf 0410}, 002 (2004)
1640:   [arXiv:hep-th/0407216].
1641:   %%CITATION = HEP-TH 0407216;%%
1642: 
1643: \bibitem{Pioline4}
1644:   B.~Durin and B.~Pioline,
1645:   ``Closed strings in Misner space: A toy model for a big bounce?,''
1646:   arXiv:hep-th/0501145.
1647:   %%CITATION = HEP-TH 0501145;%%
1648: 
1649: 
1650: \bibitem{HNP2}
1651:   Y.~Hikida, R.~R.~Nayak and K.~L.~Panigrahi,
1652:   ``D-branes in a big bang/big crunch universe: Misner space,''
1653:   arXiv:hep-th/0508003.
1654:   %%CITATION = HEP-TH 0508003;%%
1655: 
1656: \bibitem{HS}
1657:   A.~Hashimoto and S.~Sethi,
1658:    ``Holography and string dynamics in time-dependent backgrounds,''
1659:   Phys.\ Rev.\ Lett.\  {\bf 89}, 261601 (2002)
1660:   [arXiv:hep-th/0208126].
1661:   %%CITATION = HEP-TH 0208126;%%
1662: 
1663: \bibitem{AP}
1664:   M.~Alishahiha and S.~Parvizi,
1665:    ``Branes in time-dependent backgrounds and AdS/CFT correspondence,''
1666:   JHEP {\bf 0210}, 047 (2002)
1667:   [arXiv:hep-th/0208187].
1668:   %%CITATION = HEP-TH 0208187;%%
1669: 
1670: \bibitem{DN}
1671:   L.~Dolan and C.~R.~Nappi,
1672:    ``Noncommutativity in a time-dependent background,''
1673:   Phys.\ Lett.\ B {\bf 551}, 369 (2003)
1674:   [arXiv:hep-th/0210030].
1675:   %%CITATION = HEP-TH 0210030;%%
1676: 
1677: \bibitem{CLO}
1678:   R.~G.~Cai, J.~X.~Lu and N.~Ohta,
1679:   ``NCOS and D-branes in time-dependent backgrounds,''
1680:   Phys.\ Lett.\ B {\bf 551}, 178 (2003)
1681:   [arXiv:hep-th/0210206].
1682:   %%CITATION = HEP-TH 0210206;%%
1683: 
1684: \bibitem{Okuyama}
1685:   K.~Okuyama,
1686:   ``D-branes on the null-brane,''
1687:   JHEP {\bf 0302}, 043 (2003)
1688:   [arXiv:hep-th/0211218].
1689:   %%CITATION = HEP-TH 0211218;%%
1690: 
1691: \bibitem{HNP}
1692:   Y.~Hikida, R.~R.~Nayak and K.~L.~Panigrahi,
1693:   ``D-branes in a big bang/big crunch universe: Nappi-Witten gauged WZW
1694:   model,''
1695:   JHEP {\bf 0505}, 018 (2005)
1696:   [arXiv:hep-th/0503148].
1697:   %%CITATION = HEP-TH 0503148;%%
1698: 
1699: \bibitem{Hagedorn}
1700:   M.~S.~Costa, C.~A.~R.~Herdeiro, J.~Penedones and N.~Sousa,
1701:   ``Hagedorn transition and chronology protection in string theory,''
1702:   arXiv:hep-th/0504102.
1703:   %%CITATION = HEP-TH 0504102;%%
1704: 
1705: 
1706: \bibitem{MS}
1707:   J.~McGreevy and E.~Silverstein,
1708:   ``The tachyon at the end of the universe,''
1709:   arXiv:hep-th/0506130.
1710:   %%CITATION = HEP-TH 0506130;%%
1711: 
1712: \bibitem{Silverstein}
1713:   E.~Silverstein,
1714:   ``Dimensional mutation and spacelike singularities,''
1715:   arXiv:hep-th/0510044.
1716:   %%CITATION = HEP-TH 0510044;%%
1717: 
1718: \bibitem{APS}
1719:   A.~Adams, J.~Polchinski and E.~Silverstein,
1720:   ``Don't panic! Closed string tachyons in ALE space-times,''
1721:   JHEP {\bf 0110}, 029 (2001)
1722:   [arXiv:hep-th/0108075].
1723:   %%CITATION = HEP-TH 0108075;%%
1724: 
1725: 
1726: \bibitem{Martinec}
1727:   E.~J.~Martinec,
1728:   ``Defects, decay, and dissipated states,''
1729:   arXiv:hep-th/0210231.
1730:   %%CITATION = HEP-TH 0210231;%%
1731: 
1732: \bibitem{HMT}
1733:   M.~Headrick, S.~Minwalla and T.~Takayanagi,
1734:   ``Closed string tachyon condensation: An overview,''
1735:   Class.\ Quant.\ Grav.\  {\bf 21}, S1539 (2004)
1736:   [arXiv:hep-th/0405064].
1737:   %%CITATION = HEP-TH 0405064;%%
1738: 
1739: 
1740: \bibitem{She}
1741:   J.~H.~She,
1742:   ``A matrix model for Misner universe,''
1743:   arXiv:hep-th/0509067.
1744:   %%CITATION = HEP-TH 0509067;%%
1745: 
1746: 
1747: \bibitem{BKRS}
1748:   M.~Berkooz, Z.~Komargodski, D.~Reichmann and V.~Shpitalnik,
1749:   ``Flow of geometries and instantons on the null orbifold,''
1750:   arXiv:hep-th/0507067.
1751:   %%CITATION = HEP-TH 0507067;%%
1752: 
1753: 
1754: \bibitem{CSV}
1755:   B.~Craps, S.~Sethi and E.~P.~Verlinde,
1756:   ``A matrix big bang,''
1757:   arXiv:hep-th/0506180.
1758:   %%CITATION = HEP-TH 0506180;%%
1759: 
1760: \bibitem{Li}
1761:   M.~Li,
1762:   ``A class of cosmological matrix models,''
1763:   arXiv:hep-th/0506260.
1764:   %%CITATION = HEP-TH 0506260;%%
1765: 
1766: \bibitem{LS}
1767:   M.~Li and W.~Song,
1768:   ``Shock waves and cosmological matrix models,''
1769:   arXiv:hep-th/0507185.
1770:   %%CITATION = HEP-TH 0507185;%%
1771: 
1772: \bibitem{pp}
1773:   S.~R.~Das and J.~Michelson,
1774:   ``pp wave big bangs: Matrix strings and shrinking fuzzy spheres,''
1775:   arXiv:hep-th/0508068.
1776:   %%CITATION = HEP-TH 0508068;%%
1777: 
1778: \bibitem{Chen}
1779:   B.~Chen,
1780:   ``The time-dependent supersymmetric configurations in M-theory and matrix
1781:   models,''
1782:   arXiv:hep-th/0508191.
1783:   %%CITATION = HEP-TH 0508191;%%
1784: 
1785: \bibitem{RS}
1786:   D.~Robbins and S.~Sethi,
1787:   ``A matrix model for the null-brane,''
1788:   arXiv:hep-th/0509204.
1789:   %%CITATION = HEP-TH 0509204;%%
1790: 
1791: \bibitem{DDG}
1792:   D.~E.~Diaconescu, M.~R.~Douglas and J.~Gomis,
1793:   ``Fractional branes and wrapped branes,''
1794:   JHEP {\bf 9802}, 013 (1998)
1795:   [arXiv:hep-th/9712230].
1796:   %%CITATION = HEP-TH 9712230;%%
1797: 
1798: \bibitem{DG}
1799:   D.~E.~Diaconescu and J.~Gomis,
1800:   ``Fractional branes and boundary states in orbifold theories,''
1801:   JHEP {\bf 0010}, 001 (2000)
1802:   [arXiv:hep-th/9906242].
1803:   %%CITATION = HEP-TH 9906242;%%
1804: 
1805: 
1806: \bibitem{BCR}
1807:   M.~Billo, B.~Craps and F.~Roose,
1808:   ``Orbifold boundary states from Cardy's condition,''
1809:   JHEP {\bf 0101}, 038 (2001)
1810:   [arXiv:hep-th/0011060].
1811:   %%CITATION = HEP-TH 0011060;%%
1812: 
1813: \bibitem{GS}
1814:   C.~Gabriel and P.~Spindel,
1815:   ``Quantum charged fields in Rindler space,''
1816:   Annals Phys.\  {\bf 284}, 263 (2000)
1817:   [arXiv:gr-qc/9912016].
1818:   %%CITATION = GR-QC 9912016;%%
1819: 
1820: \bibitem{DM}
1821:   M.~R.~Douglas and G.~W.~Moore,
1822:   ``D-branes, Quivers, and ALE Instantons,''
1823:   arXiv:hep-th/9603167.
1824:   %%CITATION = HEP-TH 9603167;%%
1825: 
1826: \bibitem{DKPS}
1827:   M.~R.~Douglas, D.~Kabat, P.~Pouliot and S.~H.~Shenker,
1828:   ``D-branes and short distances in string theory,''
1829:   Nucl.\ Phys.\ B {\bf 485}, 85 (1997)
1830:   [arXiv:hep-th/9608024].
1831:   %%CITATION = HEP-TH 9608024;%%
1832: 
1833: \end{thebibliography}
1834: 
1835: \end{document}
1836: