1: \documentclass[twocolumn,prl,aps,floats,showpacs]{revtex4}
2: \input epsf
3: \usepackage{amssymb}
4: \begin{document}
5: %\draft
6: \title{Harmonic Measure and Winding of Conformally Invariant Curves}
7: \author{Bertrand Duplantier}
8: %\email{bertrand@spht.saclay.cea.fr}
9: \affiliation{Service de Physique Th\'{e}orique de Saclay, F-91191
10: Gif-sur-Yvette Cedex, France}
11: \author{Ilia A. Binder}
12: \affiliation{Department of Mathematics, University of Illinois at Urbana-Champaign, IL 61801}
13:
14: \date{August 2, 2002}
15:
16: \begin{abstract}
17: The exact joint multifractal distribution for the scaling
18: and winding of the electrostatic potential lines near any
19: conformally invariant scaling curve is derived in two dimensions.
20: Its spectrum $f(\alpha,\lambda)$ gives
21: the Hausdorff dimension of the points where the potential scales with distance $r$ as $H \sim r^{\alpha}$
22: while the curve logarithmically spirals with a rotation angle $\varphi=\lambda \ln r$. It obeys the scaling law
23: $f(\alpha,\lambda)=(1+\lambda^2)
24: f\left(\bar \alpha\right)-b\lambda^2$ with $\bar \alpha=\alpha/(1+\lambda^2)$ and
25: $b=(25-c)/{12}$, and where $f(\alpha)\equiv f(\alpha,0)$ is the pure
26: harmonic measure spectrum, and $c$ the conformal central
27: charge. The results apply to $O(N)$ and Potts models, as well as to
28: ${\rm SLE}_{\kappa}$.
29: \end{abstract}
30: \pacs{02.30.Em, 05.45.Df, 05.50.+q, 41.20.Cv}
31: \maketitle
32: The geometric description of the random fractals
33: arising in Nature is a fascinating subject. Among these, the
34: study of the particular class of random clusters or fractal
35: curves arising in critical phenomena has led to fundamental
36: advances in mathematical physics. In {\it two dimensions} (2D), conformal field theory (CFT)
37: has in particular demonstrated that statistical systems at their critical point
38: produce {\it conformally invariant} (CI) fractal structures,
39: examples of which are the continuum scaling limits of random
40: walks (RW), i.e., Brownian motion, self-avoiding walks (SAW), and
41: critical Ising or Potts clusters. A wealth of exact methods has
42: been devised for their study: Coulomb gas, conformal invariance, and quantum
43: gravity methods \cite{nien,cardy,DK,duplantier4,BDprl5}. Recently, rigorous methods
44: have also been developed, with the introduction of the
45: Stochastic L\"owner Evolution (${\rm SLE}$) process, which mimics the
46: wandering of critical cluster boundaries, and gives a
47: probabilistic means of studying their random geometry \cite{schramm1}.
48:
49: A refined way of accessing this random geometry is
50: provided by classical potential theory of
51: electrostatic or diffusion field near such random fractal
52: boundaries, whose self-similarity is reflected in a {\it
53: multifractal} (MF) spectrum describing the singularities of the
54: potential, also called the harmonic measure.
55: In 2D, the first exact examples appeared for the
56: universality class of random or self-avoiding walks, and
57: percolation clusters, which all possess
58: the same harmonic MF spectrum \cite{BDprl23} (see also \cite{LW}),
59: in contradistinction to higher dimensions \cite{cates}. The general solution
60: for the potential distribution near any CI fractal in 2D, obtained in \cite{BDprl5}, depends only
61: on the so-called {\it central charge c}, the parameter labeling
62: the universality class of the underlying CFT (see also \cite{LSW,hast}). This solution can be
63: generalized to higher multifractal correlations, like the joint
64: distribution of potential on both sides of a
65: simple scaling path \cite{BDjsp}.
66:
67: The important question remains of the {\it geometry
68: of the equipotential lines} near a random (CI)
69: fractal curve. They are expected to wildly rotate, or wind, in a spiralling motion which closely follows
70: the boundary itself. The key geometrical object is the {\it logarithmic spiral},
71: which is conformally invariant. The MF description should generalize to a
72: {\it mixed} multifractal spectrum, accounting for {\it both scaling
73: and winding} of the equipotentials \cite{binder}.
74:
75: In this
76: Letter, we obtain the exact solution to this mixed MF spectrum
77: for any random CI curve. In particular, it is shown
78: to be related by a scaling law to the usual
79: harmonic MF spectrum. We use conformal tools (fusing quantum
80: gravity and Coulomb gas methods), which allow the
81: description of Brownian paths interacting and winding with CI curves,
82: thereby providing a probabilistic description of the potential map.
83:
84: {\it Harmonic Measure and Rotations}. Consider a single
85: (CI) critical random cluster, generically
86: called ${\cal C}$. Let $H\left( z\right) $ be the potential at the
87: exterior point $z \in {\rm {\mathbb C}}$, with Dirichlet boundary
88: conditions $H\left({w \in \partial \cal C}\right)=0$ on the outer
89: (simply connected) boundary $\partial \cal C$ of $\cal C$, and
90: $H(w)=1$ on a circle ``at $\infty$'', i.e., of a large radius
91: scaling like the average size $R$ of $ \cal C$. As is well-known,
92: $H\left( z\right)$ is identical to the probability that a Brownian path
93: starting at $z$ escapes to ``$\infty$'' without having hit
94: ${\cal C}$.
95:
96: Let us now consider the {\it degree with which the
97: curves wind in the complex plane about point} $w$ and call
98: $\varphi(z)={\rm arg}\,(z-w)$. The multifractal formalism
99: \cite{bb,hent,frisch,halsey1}, here generalized to take into
100: account rotations \cite{binder}, characterizes subsets
101: ${\partial\cal C}_{\alpha,\lambda}$ of boundary sites by a
102: H\"{o}lder exponent $\alpha$, and a rotation rate $\lambda$,
103: such that their potential lines respectively scale and {\it logarithmic spiral} as
104: \begin{eqnarray}
105: \nonumber
106: H\left( z \to w\in {\partial\cal C}_{\alpha,\lambda }\right) &\approx& r ^{\alpha },\\
107: \varphi\left( z \to w\in {\partial\cal C}_{\alpha,\lambda
108: }\right) &\approx& \lambda \ln\, r\ , \nonumber %\label{ha}
109: \end{eqnarray}
110: in the scaling limit $a_0 \ll r=|z-w| \ll R,$ where $a_0$ is the lattice mesh, if any. The Hausdorff dimension
111: ${\rm dim}\left({\partial\cal C}_{\alpha,\lambda }\right)=f\left(
112: \alpha, \lambda\right)$ defines
113: the mixed MF spectrum, which is CI since {\it under a conformal map
114: both $\alpha$ and $\lambda$ are locally invariant}.
115:
116: Reversing the escaping Brownian path which represents the potential, one can also
117: consider the {\it harmonic measure} $H(w,r)$, which is the probability
118: that such a path starting at distance $R$ first hits the boundary in the
119: disk $B(w,r)$ of radius $r$ centered at $w \in \partial \cal C$,
120: and $\varphi (w,r)$ the associated winding angle of the path down
121: to distance $r$ from $w$. The {\it mixed} moments of $H$ and
122: $e^{\varphi}$, averaged over all realizations of ${\cal C}$, are defined as
123: \begin{equation}
124: {\cal Z}_{n,p}=\left\langle \sum\limits_{w\in {\partial {\cal C}}_r}
125: H^{n}\left(w,r\right) \exp\, (p\,\varphi (w,r)) \right\rangle
126: \approx \left( r/R\right) ^{\tau \left(
127: n,p\right) }, \label{Z}
128: \end{equation}
129: where the sum runs over the centers of a covering of the boundary by disks of radius $r$, and where $n$ and $p$ are
130: real numbers. The scaling limit involves
131: multifractal scaling exponents $\tau \left( n,p\right)$
132: which vary in a non-linear way with $n$ and $p$ \cite{binder,bb,hent,frisch,halsey1}.
133: They obey the symmetric double Legendre transform
134: \begin{eqnarray}
135: \nonumber
136: \alpha &=&\frac{\partial\tau }{\partial n}\left( n,p\right) ,
137: \quad \lambda =\frac{\partial\tau }{\partial p}\left( n,p\right), \\ \nonumber
138: f\left( \alpha, \lambda \right)&=&\alpha n+\lambda p-\tau \left( n,p\right) ,\\
139: n&=&\frac{\partial f}{\partial \alpha}\left( \alpha, \lambda
140: \right) , \quad p=\frac{\partial f}{\partial \lambda }\left(
141: \alpha, \lambda \right).
142: \label{legendre}
143: \end{eqnarray}
144: Because of the ensemble average (\ref{Z}), values of $f\left(
145: \alpha,\lambda \right)$ can become negative for some domains of
146: $\alpha,\lambda$.
147:
148: {\it Exact Mixed Multifractal Spectra}. Each 2D conformally
149: invariant random statistical system
150: can be labelled by its {\it central charge} $c$, $c\leq 1$. Our main result is the following exact scaling law:
151: \begin{eqnarray}
152: \label{scalinglaw}
153: f(\alpha,\lambda)&=&(1+\lambda^2) f\left(\frac{\alpha}{1+\lambda^2}\right)-b \lambda^2\ ,\\
154: \nonumber
155: b&\equiv&\frac{25-c}{12}\geq 2\ ,
156: \end{eqnarray}
157: where $f\left({\alpha}\right)\equiv
158: f\left({\alpha},\lambda=0\right)$ is the usual harmonic
159: MF spectrum in the absence of prescribed winding, first
160: obtained in \cite{BDprl5}, which can be recast as:
161: \begin{eqnarray}
162: f(\alpha)=\alpha+b-\frac{b\alpha^2}{2\alpha-1}\ .
163: \label{falpha}
164: \end{eqnarray}
165: We thus arrive at the very simple formula:
166: \begin{eqnarray}
167: f(\alpha,\lambda)=\alpha+b-\frac{b\alpha^2}{2\alpha-1-\lambda^2}\ .
168: \label{falphalambda}
169: \end{eqnarray}
170: Notice that by conformal symmetry ${\sup}_{\lambda}f(\alpha,\lambda)=f(\alpha,\lambda=0)$, i.e.,
171: the most likely situation in the absence of prescribed rotation
172: is the same as $\lambda=0$, i.e. {\it winding-free}.
173: The domain of definition of the usual $f(\alpha)$ (\ref{falpha}) is $\alpha \geq
174: 1/2$ \cite{BDprl5,Beur}, thus for $\lambda$-spiralling points Eq. (\ref{scalinglaw}) gives
175: \begin{eqnarray}
176: {\alpha} \geq \frac{1}{2}({1+\lambda^2})\ ,
177: \label{alpha'}
178: \end{eqnarray}
179: in agreement with a theorem by Beurling \cite{Beur,binder}.
180:
181: There is a geometrical meaning to the exponent $\alpha$. For an angle
182: with opening $\theta$, $\alpha={\pi}/{\theta}$, thus the quantity ${\pi}/{\alpha}$ can be regarded as
183: a local generalized angle with respect to the harmonic measure. The geometrical MF spectrum
184: of the boundary subset with such opening angle $\theta$ and spiralling rate
185: $\lambda$ reads from (\ref{falphalambda})
186: \begin{eqnarray}
187: \hat f(\theta,\lambda)\equiv f(\alpha=\frac{\pi}{\theta},\lambda)=\frac{\pi}{\theta}+b-b\frac{\pi}{2}
188: \left(\frac{1}{\theta}+
189: \frac{1}{\frac{2\pi}{1+\lambda^2} -\theta}\right).
190: \nonumber
191: \label{fthetalambda}
192: \end{eqnarray}
193:
194: As in (\ref{alpha'}), the domain of definition in the
195: $\theta$ variable is $0 \le \theta \le \theta(\lambda)$, with $\theta(\lambda)={2\pi}/({1+\lambda^2})$.
196: The maximum is reached when the two frontier strands about point $w$ locally collapse into a single
197: $\lambda$-spiral, whose inner opening angle is $\theta(\lambda)$ \cite{Beur}.
198:
199: In the absence of prescribed winding ($\lambda=0$), the maximum
200: $D_{\rm EP}\equiv D_{\rm EP}(0)={\sup}_{\alpha}f(\alpha,\lambda=0)$
201: gives the dimension of the {\it external perimeter} of the fractal
202: cluster, which is a {\it simple} curve without double points, and may differ from the full hull \cite{BDprl5,DAA}.
203: Its dimension reads \cite{BDprl5}
204: $D_{\rm EP}=\frac{1}{2}(1+b)-\frac{1}{2}\sqrt{b(b-2)}$.
205: This corresponds to typical values $\hat
206: \alpha=\alpha(n=0,p=0)$ and $\hat\theta={\pi}/{\hat \alpha}=\pi(3-2D_{\rm EP}).$
207:
208: For spirals, the maximum value
209: $D_{\rm EP}(\lambda)={\sup}_{\alpha}f(\alpha,\lambda)$ still
210: corresponds in the Legendre transform (\ref{legendre}) to $n=0$,
211: and gives the dimension of the {\it subset of the external
212: perimeter made of logarithmic spirals of type $\lambda$}. Owing to (\ref{scalinglaw})
213: we immediately get
214: \begin{eqnarray}
215: D_{\rm EP}(\lambda)=(1+\lambda^2)D_{\rm EP} -b \lambda^2 \ .
216: \label{supf}
217: \end{eqnarray}
218: This corresponds to scaled typical values $\hat
219: \alpha(\lambda)=(1+\lambda^2)\hat \alpha$, and $\hat
220: \theta(\lambda)=\hat \theta/(1+\lambda^2).$ Since $b \geq 2$ and
221: $D_{\rm EP} \leq 3/2$, the EP dimension decreases
222: with spiralling rate, in a simple parabolic way.
223: \begin{figure}[t]
224: \epsfxsize=8.5truecm{\centerline{\epsfbox{ff0.eps}}}
225: \caption{Universal multifractal spectrum $f(\alpha,\lambda)$ for
226: $c=0$ (Brownian frontier, percolation EP and
227: SAW), and for three different values of the
228: spiralling rate $\lambda$.} \label{Figure1}
229: \end{figure}
230: \begin{figure}[t]
231: \epsfxsize=8.5truecm{\centerline{\epsfbox{ff1.eps}}}
232: \caption{Left-sided multifractal spectra $f(\alpha,\lambda)$ for
233: the limit case $c=1$ (frontier of a $Q=4$ Potts cluster or ${\rm SLE}_{\kappa=4}$).}
234: \label{Figure2}
235: \end{figure}
236:
237: Fig. 1 displays typical multifractal functions
238: $f(\alpha,\lambda\, ;c)$. The example choosen, $c=0$,
239: corresponds to the cases of a SAW,
240: or of a percolation EP, the scaling limits of which both coincide
241: with the Brownian frontier \cite{BDprl23,LW}. The original
242: singularity at $\alpha=\frac{1}{2}$ in the
243: rotation free MF functions $f(\alpha,0)$, which describes boundary points with a needle local
244: geometry, is shifted for $\lambda \ne 0$ towards the minimal value (\ref{alpha'}). The right branch of
245: $f\left( \alpha,\lambda\right) $ has a linear asymptote
246: $\lim_{\alpha \rightarrow +\infty} f\left(\alpha,\lambda
247: \right)/{\alpha} =-(1-c)/24.$ Thus the $\lambda$-curves all become parallel for $\alpha \to
248: +\infty$, i.e., $\theta \to 0^{+}$, corresponding to deep fjords
249: where winding is easiest.
250:
251: {\it Limit} multifractal spectra are
252: obtained for $c=1$, which exhibit {\it exact} examples of {\it
253: left-sided} MF spectra, with a horizontal asymptote
254: $f\left(\alpha\to +\infty,\lambda\, ; c=1\right)=
255: \frac{3}{2}-\frac{1}{2}\lambda^2$ (Fig. 2). This corresponds to
256: the frontier of a $Q=4$ Potts cluster (i.e., the
257: ${\rm SLE}_{\kappa=4}$), a universal random scaling curve, with the
258: maximum value $D_{\rm EP}=3/2$, and a vanishing typical opening
259: angle $\hat \theta=0$, i.e., the ``ultimate Norway'' where the EP
260: is dominated by ``{\it fjords}'' everywhere
261: \cite{BDprl5,BDjsp}.
262:
263: Fig. 3 displays the dimension $D_{\rm
264: EP}(\lambda)$ as a function of the rotation
265: rate $\lambda$, for various values of $ c \leq 1$,
266: corresponding to different statistical systems. Again, the $c=1$
267: case shows the least decay with $\lambda$, as expected from the
268: predominence of fjords there.
269: \begin{figure}[t]
270: \epsfxsize=8.5truecm{\centerline{\epsfbox{Dlambda.eps}}}
271: \caption{Dimensions $D_{\rm EP}(\lambda)$ of the external frontiers as a function of rotation rate. The curves
272: are indexed by the central charge $c$, and correspond
273: respectively to: loop-erased RW ($c=-2\, ; {\rm SLE}_{2}$); Brownian or
274: percolation external frontiers, and self-avoiding walk ($c=0\, ; {\rm SLE}_{8/3}$);
275: Ising clusters ($c=\frac{1}{2}\, ; {\rm SLE}_{3}$); $Q=4$ Potts clusters ($c=1\, ; {\rm SLE}_{4}$).}
276: \label{Figure 3}
277: \end{figure}
278:
279: {\it Conformal Invariance and Quantum Gravity.} We
280: now give the main lines of the derivation of exponents $\tau\left(
281: n,p\right)$, hence $f(\alpha,\lambda),$ by generalized {\it
282: conformal invariance}. By definition of the $H$-measure, $n$ {\it
283: independent} Brownian paths ${\cal B}$, starting a small distance $r$ away from
284: a point $w$ of the frontier $\partial \cal C$, and
285: diffusing without hitting $\partial \cal C$, give a geometric
286: representation of the $n^{th}$ moment$, H^{n},$ in Eq.(\ref{Z})
287: for $n$ {\it integer}. Convexity yields analytic continuation for
288: arbitrary $n$'s. Let us introduce an abstract (conformal) field
289: operator $\Phi_{{\partial\cal C}\wedge {n}}$ characterizing the
290: presence of a vertex where $n$ such Brownian paths and the cluster's
291: frontier diffuse away from each other in a {\it mutually-avoiding} configuration noted
292: ${\partial\cal
293: C}\wedge {n}$ \cite{BDprl23}; to this operator is associated a scaling
294: dimension $x(n)$. To measure rotations as in moments (\ref{Z}) we have to consider
295: expectation values with insertion of the mixed operator
296: \begin{equation}
297: \Phi_{{\partial\cal C}\wedge n} e^{p\,{\rm arg}({\partial\cal
298: C}\wedge n)} \longrightarrow x\left( n,p\right)=\tau \left( n,p\right)+2 , \label{xnp}
299: \end{equation}
300: where ${\rm arg}({\partial\cal C}\wedge n)$ is the winding angle
301: common to the frontier and to the Brownian paths, and where $x(n,p)$ is the
302: {\it scaling dimension}. One has $x(n,p=0)=x(n)$, and
303: $\tau \left( n,p=0\right)\equiv\tau \left( n\right)=x\left(
304: n\right) -2$.
305:
306: Let us now use a fundamental mapping of the
307: CFT in the {\it plane} ${\mathbb R}^{2}$ to the CFT on a
308: fluctuating abstract random Riemann surface, i.e., in presence of
309: {\it 2D quantum gravity} (QG) \cite{KPZ}. Two universal functions $U$
310: and $V$, acting on scaling dimensions, describe this map:
311: \begin{eqnarray}
312: U\left( x\right) &=&x\frac{x-\gamma}{1-\gamma} , \hskip2mm
313: V\left( x\right) =\frac{1}{4}\frac{x^{2}-\gamma^2}{1-\gamma}.
314: \label{UV}
315: \end{eqnarray}
316: with $V\left( x\right) \equiv U\left(
317: {\textstyle{1 \over 2}}%
318: \left( x+\gamma \right) \right)$\cite{BDprl23,BDprl5}. The parameter $\gamma$
319: is the solution of
320: $c=1-6{\gamma}^2(1-\gamma)^{-1}, \gamma \leq 0.$
321:
322: For the purely harmonic exponents $x(n)$,
323: describing the mutually-avoiding set ${\partial\cal C}\wedge n$, we have \cite{BDprl23,BDprl5}
324: \begin{eqnarray}
325: x(n)&=&2V\left[2 U^{-1}\left(
326: \tilde{x_1}\right)
327: +U^{-1}\left( n \right) \right], %\nonumber \\
328: \label{xcn}
329: \end{eqnarray}
330: where $U^{-1}\left( x\right) $ is the positive inverse of $U$
331: \begin{eqnarray}
332: \nonumber
333: 2 U^{-1}\left( x\right)
334: =\sqrt{4(1-\gamma)x+\gamma^2}+\gamma\, .
335: \label{u1}
336: \end{eqnarray}
337: In (\ref{xcn}), the arguments $\tilde{x_1}$ and $n$ are respectively the {\it boundary} scaling dimensions
338: (b.s.d.) of the simple path ${\cal S}_1$ representing a
339: semi-infinite random frontier (such that ${\partial\cal C}\equiv {\cal
340: S}_1\wedge{\cal S}_1$),
341: and of the packet of $n$ Brownian paths, both diffusing into the upper {\it half-plane} $\mathbb H$.
342: The function $U^{-1}$ maps these half-plane b.s.d's to the corresponding b.s.d.'s in quantum
343: gravity, the {\it linear combination} of which gives, still in QG, the
344: b.s.d. of the {\it mutually-avoiding set} ${\partial\cal
345: C}\wedge n=(\wedge{\cal S}_1)^2\wedge n$. The function $V$ finally
346: maps the latter b.s.d. into the scaling dimension in $\mathbb R^2$.
347: The path b.s.d. $\tilde{x_1}$ obeys $U^{-1}\left(
348: \tilde{x_1}\right) =(1-\gamma)/2$ \cite{BDprl5}.
349:
350: It is now useful to consider $k$ semi-infinite
351: random paths ${\cal S}_1$, joined at a single vertex in a
352: {\it mutually-avoiding star} configuration ${\cal
353: S}_k=\stackrel{k}{\overbrace{{\cal S}_1\wedge{\cal
354: S}_1\wedge\cdots{\cal S}_1}}=(\wedge{\cal S}_1)^k$.
355: %(In this notation the frontier near any of its points is a two-star
356: %${\partial\cal C}=(\wedge{\partial\cal C}_1)^2$.
357: Its scaling dimension can be obtained from the same b.s.d. additivity rule in quantum gravity, as in
358: (\ref{xcn}) \cite{BDprl23,BDprl5}
359: \begin{eqnarray}
360: x({\cal S}_k)&=&2V\left[ k\, U^{-1}\left( \tilde{x_1}\right) \right]\ . %\nonumber \\
361: \label{xk}
362: \end{eqnarray}
363: The scaling dimensions (\ref{xcn}) and (\ref{xk}) coincide when
364: \begin{equation}
365: %\label{equal}
366: x(n)=x({\cal S}_{k(n)}),\,\,\, k(n)=2+\frac{U^{-1}\left( n
367: \right)}{U^{-1}\left( \tilde{x_1}\right)}.
368: \label{kn}
369: %\label{equal}
370: \end{equation}
371: %\begin{eqnarray}
372: %x_{\cal S}_{k(n)}}&=&x(n)\\ \ ,
373: %k(n)&=& 2+\frac{U^{-1}\left( n \right)}{U^{-1}\left( \tilde{x_1}\right)}.
374: %\label{equal}
375: %\end{eqnarray}
376: Thus we state the {\it scaling star-equivalence}
377: %\begin{eqnarray}
378: ${\partial\cal C}\wedge n \Longleftrightarrow {\cal S}_{k(n)}$,
379: %\label{equiv}
380: %\end{eqnarray}
381: {\it of two simple paths ${\cal S}_1$ avoiding $n$ Brownian motions to $k(n)$
382: simple paths in a mutually-avoiding star configuration}, an
383: equivalence which will also play a essential role in the complete rotation
384: spectrum (\ref{xnp}).
385:
386: {\it Rotation scaling exponents}. The Gaussian distribution of
387: the winding angle about the {\it extremity} of a scaling path,
388: like ${\cal S}_1$, was derived in \cite{DS}, using exact Coulomb
389: gas methods. The argument can be generalized to the winding angle
390: of a star ${\cal S}_k$ about its center \cite{BDtopub}, where
391: one finds that the angular variance is reduced by a factor
392: $1/k^2$ (see also \cite{wilson}). The
393: scaling dimension associated with the rotation scaling operator
394: $\Phi_{{\cal S}_k}e^{p\, {\rm arg }\left({\cal S}_k\right)}$ is found
395: by analytic continuation of the Fourier transforms evaluated there \cite{BDtopub}:
396: \begin{eqnarray}
397: x({\cal S}_k;p)=x({\cal
398: S}_k)-\frac{2}{1-\gamma}\frac{p^2}{k^2} \ , \nonumber %\label{xkp}
399: \end{eqnarray}
400: i.e., is given by a quadratic shift in the star scaling exponent. To calculate
401: the scaling dimension (\ref{xnp}), it is
402: sufficient to use the star-equivalence (\ref{kn}) above to
403: conclude that
404: \begin{eqnarray}
405: x(n,p)=x({\cal
406: S}_{k(n)};p)=x(n)-\frac{2}{1-\gamma}\frac{p^2}{k^2(n)} \ ,
407: \nonumber
408: \end{eqnarray}
409: which is the key to our problem. Using Eqs (\ref{kn}),
410: (\ref{xcn}), and (\ref{UV}) gives the
411: useful identity:
412: \begin{eqnarray}
413: \frac{1}{8}({1-\gamma})k^2(n)=x(n)-2+b \ ,
414: \nonumber
415: \end{eqnarray}
416: with $b=\frac{1}{2}\frac{(2-\gamma)^2}{1-\gamma}=\frac{25-c}{12}$.
417: Recalling (\ref{xnp}), we arrive at the multifractal result:
418: \begin{eqnarray}
419: \tau(n,p)=\tau(n)-\frac{1}{4}\frac{p^2}{\tau(n)+b} \ ,
420: \label{taunp}
421: \end{eqnarray}
422: where $\tau(n)=x(n)-2$ corresponds to the purely harmonic
423: spectrum with no prescribed rotation.
424:
425: {\it Legendre transform}. The structure of the full
426: $\tau$-function (\ref{taunp}) leads by a formal Legendre transform
427: (\ref{legendre}) directly to the identity
428: \begin{eqnarray}
429: f(\alpha,\lambda)&=&(1+\lambda^2)f(\bar{\alpha})-b\lambda^2 \ ,
430: \nonumber
431: \end{eqnarray}
432: where $f(\bar{\alpha})\equiv\bar{\alpha}n-\tau(n)$, with $\bar \alpha={d{\tau}(n)
433: }/{dn}$, is the purely
434: harmonic MF function. It depends on the natural reduced variable $\bar{\alpha}$
435: {\it \`a la} Beurling ($\bar{\alpha} \in \left[\frac{1}{2},+\infty \right)$)
436: \begin{eqnarray}
437: \nonumber
438: \bar{\alpha}&\equiv&\frac{\alpha}{1+\lambda^2}=\frac{d{x}
439: }{dn}\left( n\right)=\frac{1}{2} +\frac{1}{2}
440: \sqrt{\frac{b}{2n+b-2}}\, ,
441: \end{eqnarray}
442: whose expression is found explicitly from (\ref{xcn}). Whence Eq.(\ref{scalinglaw}), {\bf QED}.
443:
444: {\it $O(N)$ and Potts models, ${\rm SLE}_{\kappa}$}. Our results apply
445: to the critical $O(N)$ loop model, or to the EP's of critical Fortuin-Kasteleyn
446: (FK) clusters in the $Q$-Potts model, all described in terms
447: of Coulomb gas with some coupling constant $g$ \cite{nien}. ${\rm SLE}_{\kappa}$ paths also
448: describe cluster frontiers or hulls. One has the correspondence $\kappa={4}/{g}$, with a central charge
449: $c=(3-2g)(3-2g')=\frac{1}{4}({6-\kappa})(6-{\kappa'})$,
450: symmetric under the {\it duality} $gg'=1$ or $\kappa \kappa'=16$. This duality gives FK-EP's as
451: some simple random $O(N)$ loops, or, equivalently, the ${\rm SLE}_{\kappa'\leq 4}$ as the simple frontier of the
452: ${\rm SLE}_{\kappa \geq 4}$ \cite{BDprl5,BDjsp}.
453:
454: \begin{acknowledgments}
455: The gracious hospitality of the Mittag-Leffler Institute, where
456: this work was initiated, is gratefully acknowledged, as is a
457: careful reading of the manuscript by T. C. Halsey.
458: \end{acknowledgments}
459:
460: \begin{thebibliography}{99}
461: \bibitem{nien}
462: see, e.g., B. Nienhuis, in {\it Phase Transitions and Critical Phenomena}, edited by C.
463: Domb and J. L. Lebowitz, (Academic, London, 1987), Vol. 11.
464:
465: \bibitem{cardy} J. C. Cardy, {\it Lectures on Conformal Invariance and Percolation},
466: Chuo University, (2001).
467:
468: \bibitem{DK} B. Duplantier and I.K. Kostov, Nucl. Phys. {\bf B340}, 491 (1990).
469:
470:
471: \bibitem{duplantier4} B. Duplantier, Phys. Rev. Lett. {\bf 81}, 5489 (1998).
472:
473: \bibitem{schramm1} O. Schramm, Israel Jour. Math., {\bf 118}, 221 (2000).
474:
475: \bibitem{BDprl23} B. Duplantier, Phys. Rev. Lett. {\bf 82}, 880, 3940 (1999).
476:
477: \bibitem{LW} G.F. Lawler and W. Werner, Ann. Probab. {\bf 27}, 1601 (1999) ;
478: J. Eur. Math. Soc. {\bf 2}, 291 (2000)\, ; J. Cardy, J. Phys. A
479: {\bf 32}, L177 (1999).
480:
481: \bibitem{cates} M.E. Cates and T.A. Witten, Phys. Rev. A {\bf 35}, 1809 (1987).
482:
483: \bibitem{BDprl5} B. Duplantier, Phys. Rev. Lett. {\bf 84}, 1363 (2000).
484:
485: \bibitem{LSW} G.F. Lawler, O. Schramm and W. Werner, Acta Math. {\bf 187}, 275 (2001).
486:
487: \bibitem{hast} M. B. Hastings, Phys. Rev. Lett. {\bf 88}, 055506 (2002).
488:
489: \bibitem{BDjsp} B. Duplantier, J. Stat. Phys., to appear, cond-mat/0207743.
490:
491:
492: \bibitem{binder} I. A. Binder, {\it Harmonic Measure and Rotation of Simply Connected Planar Domains},
493: PhD Thesis, Caltech (1998).
494:
495: \bibitem{bb} B.B. Mandelbrot, J. Fluid. Mech. {\bf 62}, 331 (1974).
496:
497: \bibitem{hent} H.G.E. Hentschel and I. Procaccia, Physica (Amsterdam) {\bf %
498: 8D}, 835 (1983).
499:
500: \bibitem{frisch} U. Frisch and G. Parisi, in Proceedings of the
501: International School of Physics ``Enrico Fermi'', course
502: LXXXVIII, edited by M. Ghil (North-Holland, New York, 1985) p. 84.
503:
504: \bibitem{halsey1} T.C. Halsey, M.H. Jensen, L.P. Kadanoff, I. Procaccia,
505: and B.I. Shraiman, Phys. Rev. A {\bf 33}, 1141 (1986).
506:
507: \bibitem{Beur}
508: A. Beurling, \emph{The collected works of {A}rne {B}eurling. {V}ol.
509: 1},
510: Contemporary Mathematicians, Birkh\"auser Boston Inc., Boston, MA,
511: 1989,
512: Complex analysis, Edited by L. Carleson, P. Malliavin, J. Neuberger
513: and J.
514: Wermer.
515:
516: \bibitem{DAA} M. Aizenman, B. Duplantier, and A. Aharony, Phys. Rev. Lett. {\bf 83}, 1359 (1999).
517:
518: \bibitem{KPZ} V.G. Knizhnik, A.M. Polyakov, and A.B. Zamolodchikov, Mod.
519: Phys. Lett. A {\bf 3}, 819 (1988).
520:
521: \bibitem{DS}
522: B. Duplantier and H. Saleur, Phys. Rev. Lett. {\bf 60}, 2343
523: (1988).
524:
525: \bibitem{BDtopub}
526: B. Duplantier, to be published.
527:
528: \bibitem{wilson}
529: B. Wieland and D. B. Wilson, to be published.
530:
531: \end{thebibliography}
532:
533: \end{document}
534:
535:
536: