Critical exponents, conformal invariance and planar Brownian motion (Q4407972)

This review-type paper focuses on the scaling limit of 2d-systems and conformally invariant objects as candidates for the limit. The author describes first some conjectures of physicists from the eighties concerning the asymptotic behavior of self-avoiding walks and critical planar percolation. These predictions heavily rely on the common belief that the scaling limit of 2d-systems in statistical physics exhibits conformally invariant behavior at criticality. For self-avoiding random walks, the underlying belief is that the uniform measure on long self-avoiding walks, suitably rescaled, converges to a measure on continuous curves that possesses some invariance properties under conformal transformations. Concerning critical planar percolation, physicists predict that the geometry is lattice-independent and is also invariant under conformal transformations in the scaling limit. Exact formulas concerning the limit have even been produced and the author recalls here a prediction concerning the Hausdorff dimension of the limit law of the cluster boundaries and others related to the famous Cardy's formula, giving the crossing probabilities of a quadrilateral. The second part of this review paper describes results (from the author and its co-authors) that prove some conjectures and improve the understanding of others. The first theorem states that the probability that two independent Brownian paths on \([0,t]\) starting at the origin are disjoints decays in the limit like \(t^{-5/8}\), a result predicted 15-20 years ago by Duplantier and Kwon. The second result proves Mandelbroot's conjecture on planar Brownian motion: almost surely, the Hausdorff dimension of the outer boundary of the hull of \(B[0,1]\) is \(4/3\). Some other consequences on Hausdorff dimensions of particular sets are also deduced. The author then describes a notion of universality crucial in the proof of the previous results, expressed in terms of ``completely conformally invariance'' (CCI). Roughly speaking, a CCI measure is a measure on continuous curves invariant under conformal transformations with restrictions on some smooth domains. Afterwards, the author states results on the existence and uniqueness of CCI probability measure in the plane and in the half-plane (with a relation with Cardy's formula) and describes also the construction of CCI probability measures based on Loewner's differential equations that encodes a certain class of growing families of compact sets. In a particular case (SLE\(_6\)), the construction in the plane (radial Schramm's process) and in the half-plane (chordal Schramm's process) share the same local evolution and can be expressed in terms of reflected Brownian motions. The author sketches then how SLE\(_6\) can be used to compute Brownian exponents described above. In the last section, the author describes Schramm's conjecture which identifies SLE\(_6\) with the scaling limit of the so-called percolation-exploration process, a first step towards the identification of the scaling limit of critical percolation. This review paper is a nice introduction to the state of art of this difficult and active subject, just before major improvements made after the publication of this paper. For the entire collection see [Zbl 0972.00032].