1: \documentclass[twocolumn,prl]{revtex4}
2: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
3: \newcommand\p{\partial}
4: \newcommand\CD{{\cal D}}
5: \newcommand\CT{{\cal T}}
6: \newcommand\CV{{\cal V}}
7: \newcommand\be{\begin{equation}}
8: \newcommand\ee{\end{equation}}
9: \newcommand\bea{\begin{eqnarray}}
10: \newcommand\eea{\end{eqnarray}}
11: \renewcommand\tilde[1]{\widetilde{#1}}
12: \renewcommand\hat[1]{\widehat{#1}}
13: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
14: \begin{document}
15: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
16: \title{Closed-String Tachyon Condensation and the Worldsheet
17: Super-Higgs Effect}
18: %
19: \author{Petr Ho\v rava and Cynthia A. Keeler}
20: %
21: \affiliation{Berkeley Center for Theoretical Physics and Department of
22: Physics,\\
23: University of California, Berkeley, California 94720-7300\\
24: and\\
25: Physics Division, Lawrence Berkeley National Laboratory,\\
26: Berkeley, California 94720-8162, USA}
27: %
28: \begin{abstract}
29: Alternative gauge choices for worldsheet supersymmetry can elucidate dynamical
30: phenomena obscured in the usual superconformal gauge. In the particular
31: example of the tachyonic $E_8$ heterotic string, we use a judicious gauge
32: choice to show that the process of closed-string tachyon condensation can be
33: understood in terms of a worldsheet super-Higgs effect. The worldsheet
34: gravitino assimilates the goldstino and becomes a dynamical propagating
35: field. Conformal, but not superconformal, invariance is maintained throughout.
36: \end{abstract}
37: \pacs{}
38: \maketitle
39: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
40: In theories with local worldsheet supersymmetry, it is traditional to work in
41: superconformal gauge, sometimes supplemented by a subsidiary gauge-fixing
42: condition such as light-cone gauge. This superconformal gauge choice has been
43: successful in a large class of static backgrounds, but it may obscure aspects
44: of more complicated systems.
45:
46: String theory has now reached the stage which requires answers to a new class
47: of dynamical questions, involving backgrounds far from equilibrium and with
48: substantial time dependence. Such questions are particularly pressing
49: in the cosmological setting. This challenge is likely to require new
50: techniques.
51:
52: We propose using a meaningful alternative to superconformal gauge; this
53: previously unexplored tool can lead to new insights into string dynamics.
54: To illustrate this phenomenon, we focus on a particular
55: example which epitomizes time-dependent processes: The problem of
56: closed-string tachyon condensation. While understanding the dynamics of its
57: open-string cousins has improved our picture of brane decay into the vacuum
58: (or to lower-dimensional stable branes), the closed-string tachyon is related
59: to the possible decay of spacetime itself. This makes the problem difficult,
60: and highlights the current
61: limitations in our undestanding of string theory far from equilibrium.
62:
63: In this letter, we demonstrate the use of alternative worldsheet gauge choices
64: in the specific example of a ten-dimensional heterotic string model with $E_8$
65: gauge symmetry and a singlet tachyon. This little-studied model belongs to the
66: early classification of modular invariant string theories in ten
67: dimensions \cite{klt}, but its {\it raison d'\^etre\/} has been hitherto
68: obscure. In a more detailed companion paper \cite{pixie}, we argue that
69: this $E_8$ heterotic string describes the fate of an unstable
70: configuration in heterotic M-theory with the two $E_8$ boundaries
71: breaking complementary sets of sixteen supercharges \cite{fh}. This
72: configuration suffers from the ``decay to nothing'' instability
73: (at least in supergravity), due to an instanton connecting the two
74: boundaries by a throat. Since there is an attractive Casimir force
75: between the two $E_8$ boundaries, the fate of the ``decay-to-nothing''
76: instability should be addressed at weak string coupling where it turns
77: into a perturbative tachyonic instability. The unique tachyonic $E_8$
78: string is the only promising candidate for describing this decay.
79:
80: In the present work we put this motivation aside, and study the
81: tachyon decay in the weakly coupled heterotic $E_8$ model as an
82: independently interesting problem.
83: The worldsheet theory of the heterotic $E_8$ string consists of $(0,1)$ chiral
84: supergravity, coupled to matter supermultiplets. In the fermionic
85: representation, the worldsheet action of the $E_8$ string is locally identical
86: to that of its supersymmetric cousins,
87: %
88: \bea
89: S_0&=&\frac{1}{4\pi\alpha'}\int d^2\sigma e\left\{(h^{mn}\p_m X^\mu\p_n X^\nu
90: +\frac{i}{2}\psi^\mu\gamma^mD_m\psi^\nu\right.\nonumber\\
91: &&{}+i\chi_m\gamma^n\gamma^m\p_nX^\mu\psi^\nu)\eta_{\mu\nu}
92: +\frac{i}{2}\lambda^A\gamma^mD_m\lambda^A\nonumber\\
93: &&\left.\vphantom{\frac{1}{2}}\qquad\qquad{}+F^AF^A\right\}.
94: \eea
95: %
96: The matter fermions are chiral, with $\Gamma\lambda=\lambda$,
97: $\Gamma\psi=-\psi$ where $\Gamma\equiv\gamma^0\gamma^1$, and the gravitino
98: satisfies $\Gamma\chi_m=\chi_m$. We are in Minkowski signature, with
99: $h_{mn}=e_m{}^ae_n{}^b\eta_{ab}$.
100:
101: This tachyonic $E_8$ model is distinguished by its GSO projection:
102: It is the only one in which no two fermions among $\lambda^A$ carry
103: always the same spin structure. It turns out that 31 of the
104: $\lambda^A$s give rise to the level-two $E_8$ current algebra, of
105: left-moving central charge $c_L=31/2$. We denote by $A'$ the index
106: that runs over those first 31 values of $A$. The remaining $1/2$ of
107: central charge is supplied by the remaining fermion
108: $\lambda^{32}\equiv\lambda$. The spin structure of $\lambda$ is
109: always the same as that of the right-moving NSR fermions $\psi^\mu$,
110: a property that singles out $\lambda$ uniquely. As we shall see, this
111: ``lone fermion'' $\lambda$ plays a central role in the process of
112: tachyon condensation.
113:
114: We wish to examine the spacetime evolution of the tachyon condensate
115: as an on-shell process in string theory. Thus our worldsheet theory
116: must maintain conformal invariance. Accordingly, we turn on a
117: linear dilaton background $\Phi=V_\mu X^\mu$. Since we do not wish
118: to change the critical dimension, the dilaton gradient $V$ is chosen
119: to be null. We choose spacetime lightcone coordinates $X^\pm=X^0\pm
120: X^1$ (with the spacetime index $\mu$ splitting as $(\pm,i)$, $i=1,\ldots 8$),
121: such that the only nonzero component of $V$ is $V_-$. The action is
122: $S_0+S_V$, where
123: %
124: \be
125: S_V=-\frac{1}{4\pi}\int d^2\sigma e\,V_-\left(X^-R+2i\chi_m\gamma^n
126: \gamma^mD_n\psi^-\right).
127: \ee
128: %
129: This is invariant under local supersymmetry transformations, given by
130: %
131: \bea
132: \delta e_m{}^a&=&2i\epsilon\gamma^a\chi_m,\quad
133: \delta\chi_m=D_m\epsilon\nonumber\\
134: \delta X^\mu&=&i\epsilon\psi^\mu,\quad\delta\psi^\mu=\gamma^m\p_m X^\mu\epsilon
135: +\alpha'V^\mu\gamma^m D_m\epsilon,\nonumber\\
136: \delta\lambda^A&=&F^A\epsilon,\quad\delta F^A=i\epsilon\gamma^mD_m\lambda^A.
137: \label{susyrules}
138: \eea
139: %
140: At this stage, the auxiliary fields $F^A$ are usually integrated out;
141: we indeed set $F^{A'}=0$, but keep $F\equiv F^{32}$, as it plays a rather
142: notrivial role in the dynamics of the tachyon condensation. Indeed, the
143: tachyon vertex operator with spacetime momentum $p_\mu$ is related to the lone
144: fermion supermultiplet, and given (in picture~0) by
145: %
146: \be
147: \CV=\left(F+\lambda\,p_\mu\psi^\mu\right)\exp(ip_\mu X^\mu).
148: \ee
149: %
150:
151: We now turn on a condensate of the tachyon in the form of worldsheet
152: superpotential
153: %
154: \be
155: \label{supp}
156: S_W=-\frac{\mu}{\pi\alpha'}\int d^2\sigma\left\{F\CT(X)-i\lambda
157: \psi^\mu\p_\mu\CT(X)\right\},
158: \ee
159: %
160: and choose $\CT(X)$ such that (\ref{supp}) is an exactly marginal
161: deformation in conformal gauge. This requires
162: %
163: \be
164: \CT(X)=\exp(k_+X^+)\quad {\rm and}\quad -k^2+2V\cdot k=\frac{2}{\alpha'}.
165: \ee
166: %
167: The real constant $k_+>0$ is thus related to the dilaton gradient via
168: %
169: \be
170: V_-k_+=-\frac{1}{2\alpha'}.
171: \ee
172: %
173: These conditions arise at the first and second order of conformal
174: perturbation theory in the dimensionless $\mu$.
175:
176: As the next step, we need to choose a gauge. In critical string
177: theories with worldsheet supersymmetry, superconformal gauge -- in
178: which $e_m{}^a=\delta_m{}^a$ and the gravitino $\chi_m$ is set to zero -- is
179: virtually always chosen. The residual superconformal
180: symmetry of this gauge is sometimes further fixed by going additionally
181: to lightcone gauge; however, this extra gauge fixing is supplemental to
182: the conformal gauge and not an alternative to it.
183:
184: To motivate a meaningful alternative gauge, consider the transformation
185: properties of the lone fermion $\lambda$ in (\ref{susyrules}). In the
186: presence of the tachyon condensate, $F$ develops a vacuum expectation value,
187: and can be integrated out using its algebraic equation of motion,
188: %
189: \be
190: \label{cond}
191: F=2\mu\exp(k_+X^+).
192: \ee
193: %
194: In the rest of this letter, we use $F$ as a shorthand for the
195: composite operator (\ref{cond}).
196:
197: In the presence of the condensate, the supersymmetry transformation
198: $\delta\lambda=F\epsilon$ indicates that $\lambda$ may play the role of
199: a goldstino for spontaneously broken worldsheet supersymmetry. Since
200: this symmetry is local, this in turn suggests that a super-Higgs mechanism
201: takes place, in which the gravitino develops a propagating degree of freedom.
202: As a result, it may be wise not to set the gravitino to zero by gauge choice,
203: and an alternative to the superconformal gauge is required.
204:
205: In higher dimensions, the Higgs mechanism is usually associated with
206: the gauge field becoming massive. In our case, no mass scale should
207: be generated if the worldsheet theory is to stay conformal. In our
208: view, the more general signature of the Higgs mechanism is that the
209: gauge field acquires more physical polarizations. In four spacetime
210: dimensions, this happens if a massless gauge field develops a mass.
211: The worldsheet situation is different. In the absence of tachyon
212: condensation, the worldsheet gravitino imposes a constraint, setting
213: the supercurrent to zero on physical states; it effectively carries
214: ``minus one'' polarization. When the tachyon condensate develops,
215: the worldsheet gravitino will acquire a propagating polarization:
216: The net number of polarizations increases, but -- unlike in four
217: dimensions -- no mass is acquired.
218:
219: Our alternative gauge for worldsheet supersymmetry still involves fixing
220: worldsheet diffeomorphisms by the conventional conformal gauge,
221: %
222: \be
223: \label{cg}
224: e_m{}^a=\delta_m{}^a.
225: \ee
226: %
227: This is consistent with our expectation of worldsheet conformal invariance.
228: From now on, our analysis continues in this gauge; and we use worldsheet
229: lightcone coordinates $\sigma^\pm=\tau\pm\sigma$. In these coordinates, the
230: only nonzero components of the fermi fields are $\chi_{++}$, $\psi_-^\mu$,
231: and $\lambda_+^A$.
232:
233: Since the gravitino is expected to gain a physical polarization at
234: the expense of the goldstino $\lambda$, the first instinct for a
235: better gauge might be to set $\lambda=0$. The equation of motion
236: that follows from varying $\lambda$ would then be imposed as a
237: constraint on physical states. In our case, this would imply
238: %
239: \be
240: \mu\psi_-^+\exp(k_+X^+)=0,
241: \ee
242: %
243: which can be solved by setting $\psi_-^+=0$. This would in turn require that
244: the equation of motion obtained from varying $\psi_-^+$ becomes a new
245: constraint.
246:
247: The benefit of such an alternative gauge choice is that the gravitino is not
248: artificially set to zero, and is free to develop its own dynamics.
249: Once we set $\psi_-^+=0$ in the full action, as tenatively suggested by our
250: first look at this alternative gauge above, the gravitino couples to the rest
251: of the system in a relatively simple way, via the terms
252: %
253: \be
254: \label{terms}
255: \chi_{++}\psi_-^-\p_-X^++2\alpha'V_-\chi_{++}\p_-\psi_-^-
256: -2\chi_{++}\psi^i\p_-X^i
257: \ee
258: %
259: in the action. This is further simplified by rescaling
260: %
261: \be
262: \label{resgrav}
263: \tilde\chi_{++}=F\chi_{++},\qquad\tilde\psi_-^-=\frac{\psi_-^-}{F}.
264: \ee
265: %
266: This rescaling changes the conformal weights of these fields from $(1,-1/2)$
267: and $(0,1/2)$ to $(3/2,0)$ and $(-1/2,0)$ respectively, and transforms
268: (\ref{terms}) to
269: %
270: \be
271: 2\alpha'V_-\tilde\chi_{++}\p_-\tilde\psi_-^--\frac{2}{F}
272: \tilde\chi_{++}\psi^i\p_-X^i.
273: \ee
274: %
275: Thus, $\tilde\chi_{++}$ and $\tilde\psi_-^-$ appear to form a left-moving
276: first-order system of spin $3/2$ and $-1/2$ and central charge $c_L=-11$,
277: whose coupling to the transverse degrees of freedom $X^i, \psi^i$ becomes
278: arbitrarily weak at late $X^+$. This left-moving first-order system is
279: our dynamical gravitino sector!
280:
281: In the precise implementation of this alternative gauge, a slight modification
282: of the process outlined above is useful. Thus, as our gauge choice, we choose
283: to supplement (\ref{cg}) by setting
284: %
285: \be
286: \label{pixiegauge}
287: \psi_-^+=0,
288: \ee
289: %
290: instead of setting $\lambda_+=0$ (or $\chi_{++}=0$ as in superconformal
291: gauge). As we will see, this accomplishes all the expected benefits outlined
292: above, and leads to other simplifications.
293:
294: The Faddeev-Popov superdeterminant $\Delta_{\rm FP}$ for our gauge choice
295: factorizes, $\Delta_{\rm FP}=J_{bc}/J_{\psi_-^+\epsilon}$. Here $J_{bc}$ is
296: the determinant associated with the conformal gauge for worldsheet
297: diffeomorphisms, traditionally realized by the path integral over a $bc$
298: ghost system of spin 2 and central charge $c_L=c_R=-26$.
299: $J_{\psi_-^+\epsilon}$ is the determinant of the operator associated with the
300: change of variables in the fermionic sector, from $\psi_-^+$ to $\epsilon$,
301: as determined from the supersymmetry variation of the gauge-fixing condition,
302: %
303: \bea
304: \label{trans}
305: \delta\psi^+&=&\gamma^m\p_mX^+\epsilon-2\alpha'V_-\gamma^mD_m\epsilon
306: \nonumber\\
307: &=&\frac{4\alpha'V_-}{F}D_-(F\epsilon).
308: \eea
309: %
310: Hence, $J_{\psi_-^+\epsilon}$ is the determinant of
311: %
312: \be
313: \label{oper}
314: D_F\equiv \frac{1}{F}D_-F,
315: \ee
316: %
317: acting on fields of spin $j=1/2$. When properly regularized and
318: renormalized, this determinant will depend not only on the Liouville
319: field $\phi$ but also on $X^+$ due to the explicit appearance of
320: $F$ in (\ref{trans}). This determinant can be explicitly evaluated
321: using formulas found for example in \cite{kallosh} (see \cite{pixie}
322: for details). It is useful to consider the slight generalization of
323: $D_F$ given by (\ref{oper}) as acting on fields of spin $j$; the
324: determinant of the corresponding Laplacian is
325: %
326: \bea
327: \label{FPdet}
328: &&\frac{1}{2}\log\,\det(D^\dagger_F D_F)\nonumber\\
329: &&{}=-\frac{1}{48\pi}\int d^2\sigma\,\hat e\left\{[3(2j-1)^2-1]
330: \hat h^{mn}\p_m\phi\p_n\phi\right.\nonumber\\
331: &&\qquad\qquad +12(2j-1)k_+\hat h^{mn}\p_m\phi\p_n X^+\nonumber\\
332: &&\quad\qquad\qquad\qquad\left.{}+12k_+^2\hat h^{mn}\p_mX^+\p_nX^+\right\},
333: \eea
334: %
335: where $h_{mn}=e^{2\phi}\hat h_{mn}$, and $\hat h_{mn}$ is a fiducial metric.
336:
337: At $j=1/2$, the first term can be realized by introducing a bosonic
338: ghost-antighost pair, $\tilde\beta$, $\tilde\gamma$. These are
339: standard left-moving first-order fields of conformal weight
340: $(1/2,0)$. The second term in (\ref{FPdet}) vanishes for $j=1/2$,
341: and the third, $X^+$ dependent term is a one-loop contribution to
342: the action of the $X^\pm$ system. Note that the ghosts $\tilde\beta$,
343: $\tilde\gamma$ associated with our gauge choice are purely
344: {\it left-moving\/}, and contribute $c_L=-1$ to the {\it bosonic\/} sector
345: on the worldsheet. This is to be contrasted with conventional superconformal
346: ghosts in superconformal gauge, which are {\it right-moving\/} and contribute
347: $c_R=11$ to the supersymmetric sector instead.
348:
349: In our alternative gauge, the worldsheet theory thus looks very different
350: than in superconformal gauge. We have gained a left-moving propagating
351: massless gravitino of spin 3/2, which (together with its conjugate partner
352: $\tilde\psi_-^-$ of spin $-1/2$) contributes central charge $c_L=-11$.
353: We have also gained left-moving bosonic ghosts of spin 1/2, contributing an
354: additional $-1$ to $c_L$; altogether we have lost twelve units of left-moving
355: central charge. Similarly, in the right-moving sector we have lost
356: the $\psi_-^\pm$ pair of fermions, worth $c_R=1$: $\psi_-^+$ is zero
357: by our gauge choice, and $\psi_-^-$ has turned into a left-moving
358: field. In addition, no conventional right-moving superconformal
359: ghosts $\beta, \gamma$ of central charge $c_R=11$ arise, leading to
360: the same apparent total discrepancy of 12 units of central charge as
361: in the left-moving sector.
362:
363: Does this mean that the theory lost its conformal invariance? The answer
364: is no: A careful analysis of the changes of variables ({\it i.e.}, the gauge
365: fixing and the $X^+$ dependent rescaling of fields) reveals one-loop
366: corrections due to the measure factors that precisely shift the dilaton from
367: its originally null direction, making up for the missing twelve units of
368: central charge and restoring quantum conformal invariance. The one-loop
369: correction from the Faddeev-Popov determinant to the effective action
370: has been calculated above, and is given by
371: %
372: \be
373: \label{xx}
374: -\frac{1}{4\pi}\int d^2\sigma\,\hat e\,k_+^2\hat h^{mn}\p_mX^+\p_nX^+.
375: \ee
376: %
377: Another similar contribution comes from the rescaling of the fields in the
378: gravitino sector (\ref{resgrav}). This change has led to a canonical kinetic
379: term in the classical action, but it did not produce the canonical measure
380: for such first-order fields: The original measure of the $\chi_{++}$ and
381: $\psi_-^+$ fields maps to a measure with an explicit $X^+$ dependence for the
382: rescaled fields. We wish to replace this $X^+$ dependent measure by the
383: canonical one. This generates a conversion factor, which can be evaluated
384: by comparing the Gaussian part of the path integral in the original
385: $\chi_{++}, \psi_-^-$ variables, to the path integral in the new variables
386: $\tilde\chi_{++}, \tilde\psi_-^-$ {\it with respect to the canonical measure
387: independent of $X^+$}. The former is obtained from (\ref{FPdet}) for
388: $j=3/2$, and is given by
389: %
390: \bea
391: \label{gravdet}
392: &&\log\int\CD\chi_{++}\,\CD\psi_-^-\,e^{-S_2(\chi_{++},\psi_-^+,X^+)}
393: \nonumber\\
394: &&\qquad{}=-\frac{1}{48\pi}\int d^2\sigma\,\hat e\left\{11\hat h^{mn}\p_m\phi
395: \p_n\phi\right.\\
396: &&\left.{}+24k_+\hat h^{mn}\p_m\phi\p_n X^++12k_+^2\hat h^{mn}\p_mX^+
397: \p_nX^+\right\},\nonumber
398: \eea
399: %
400: where $S_2$ is the part of the action bilinear in $\chi_{++}$ and
401: $\psi_-^-$. The integral over $\tilde\chi_{++}$ and $\tilde\psi_-^-$
402: (with the measure independent of $X^+$) reproduces correctly the
403: Weyl anomaly represented by the first, Liouville-dependent term; this
404: confirms that these fields represent a left-moving first-order system with
405: $c_L=-11$. The remaining two terms in (\ref{gravdet}) are the conversion
406: factor due to the replacement of the measure in the gravitino sector by the
407: canonical one. The third term cancels the $X^+$ dependent
408: contribution (\ref{xx}) of the Faddeev-Popov determinant. The second
409: one then shifts the linear dilaton term in the action (which we again
410: write in Minkowski signature), by
411: %
412: \be
413: \label{loopdil}
414: \Delta S=-\frac{1}{4\pi}\int d^2\sigma\, e\,k_+X^+R.
415: \ee
416: %
417: Here we have integrated the $\p\phi\p X^+$ term in (\ref{gravdet})
418: by parts, and used the expression $e R=-2\hat e\hat\nabla^2\phi$ for
419: the scalar curvature in conformal gauge. The one-loop correction
420: (\ref{loopdil}) makes the dilaton spacelike, and provides the missing
421: $c=12$ units of the central charge. The worldsheet theory is conformal at
422: the quantum level.
423:
424: The proper implementation of the BRST quantization involves the BRST partner
425: $B$ of the antighost $\tilde\beta$. The integration over $B$ produces a
426: delta function which imposes the gauge fixing condition. Once $B$ is
427: integrated out, it needs to be eliminated from the BRST transformation rule
428: for $\tilde\beta$. The new BRST variation of $\tilde\beta$ gives essentially
429: the equation of motion resulting from the variation of $\psi_-^+$, and must
430: vanish on physical states. This constraint has a clear intuitive
431: meaning: It takes the form $\sim\mu\lambda_+\exp(k_+X^+)=\ldots$, where the
432: $\lambda_+$ term comes from the variation of the superpotential, and
433: ``$\ldots$'' are the contributions from all the other terms, none of which
434: depends on $\lambda_+$. This simply means that the goldstino $\lambda_+$ is
435: not a separate dynamical field, but can be solved for in terms of the
436: other fields, including the now-dynamical gravitino. This is how the
437: super-Higgs mechanism is implemented in this gauge.
438:
439: Note also that our gauge fixing condition leaves no residual supersymmetry
440: at finite $\mu$. Thus, in our alternative gauge, the gauge-fixed theory is
441: not superconformal, even though manifest conformal invariance is maintained.
442:
443: The picture of the heterotic $E_8$ model in our alternative gauge can be
444: easily extended to other models with worldsheet supersymmetry. For example,
445: in Type 0 string theory in the presence of null linear dilaton $\Phi=V_-X^-$
446: and an exponential tachyon condensate along $X^+$ (studied extensively in
447: \cite{hs3}), we can choose a gauge by setting both $\psi_-^+$ and $\psi_+^+$
448: (the left-moving and right-moving superpartners of $X^+$) to zero. Compared
449: to the heterotic model, we expect the entire gauge-fixing procedure to be
450: doubled, resulting in spin 1/2 bosonic ghosts $\tilde\beta$, $\tilde\gamma$
451: and a propagating gravitino sector $\tilde\chi$, $\tilde\psi^-$ in the
452: left-moving as well as right-moving sector. Before the one-loop effects of
453: the Faddeev-Popov determinant and of the rescaling of the gravitino sector are
454: incorporated, the apparent discrepancy in the central charge is $-24$ for each
455: movers. The one-loop determinants then shift the dilaton by twice the
456: heterotic amount, adjusting the central charge by $24$ and restoring exact
457: conformal invariance at the quantum level. In the process, the worldsheet
458: gravitino has again been liberated, leading to a free fermion system of
459: $c=-11$. This elucidates some of the results of \cite{hs3} (as well as the
460: related construction of \cite{bv}), in which a fermionic $c=-11$ system was
461: found in superconformal gauge. Our gauge shows that this system shoud be
462: interpreted as the dynamical worldsheet gravitino.
463:
464: The two examples of tachyon condensation considered above suggest that
465: alternative worldsheet gauge choices may illuminate a more general
466: class of string backgrounds out of equilibrium. Yet, the physics of
467: observables in any given background should be independent of the
468: choice of gauge, suggesting a new type of worldsheet duality in
469: string theory: A duality between different gauge choices of the
470: same worldsheet theory. This clearly extends the class of
471: perturbative string dualities encountered so far: Instead of
472: representing one CFT in two dual ways, we can now look at one string
473: background in two gauges, in which they are not even described by
474: the same CFT. This feature appears to be conceptually new, and we
475: hope that it will be useful in understanding string backgrounds with
476: substantial time dependence, beyond the example of heterotic tachyon
477: condensation studied here.
478:
479: This work has been supported by NSF Grants PHY-0244900 and PHY-0555662, DOE
480: Grant DE-AC03-76SF00098, and the Berkeley Center for Theoretical Physics.
481: One of us (C.A.K.) has also been supported by an NSF Graduate Fellowship.
482: \bibliography{8}
483: \end{document}
484: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
485: