Multifractality of wave functions at the quantum Hall transition revisited

Abstract

We investigate numerically the statistics of wave function amplitudes (r) at the integer quantum Hall transition. It is demonstrated that in the limit of a large system size the distribution function of ͉͉ 2 is log-normal, so that the multifractal spectrum f (␣) is exactly parabolic. Our findings lend strong support to a recent conjecture for a critical theory of the quantum Hall transition.In 1980 von Klitzing, Dorda, and Pepper discovered 1 that the Hall conductance xy of a two-dimensional electron gas develops plateaus at values quantized in units of e 2 /h. Twenty years later, the integer quantum Hall effect ͑IQHE͒ still constitutes one of the great challenges of condensedmatter physics. Initially, the effort focused on the physics of the Hall plateaus which is by now fairly well understood. 2,3 Then interest has shifted towards the transition region, where xy crosses over from one plateau to the next. However, here the situation is not as well resolved. It has been understood from early on that this is a second-order phase transition, and the scaling scenario has been confirmed in numerous experiments and computer simulations 4 yielding, e.g., the value Ӎ2.35 for the localization length exponent. In contrast, analytical approaches did not lead to quantitative predictions. The ultimate goal here is to identify the effective low-energy theory of the critical point ͑expected to be a conformal field theory͒ and to calculate critical exponents and other characteristics of the transition region. In particular, independence of the critical theory on microscopic parameters would establish universality of the critical exponents which is a matter of intensive and controversial discussions in the experimental literature.The earliest field-theoretical formulation of the problem was given by Pruisken 5 ͑see Ref. 6 for a more precise, supersymmetric version͒ and has the form of the nonlinear -model with a topological term. Since the latter is invisible in perturbation theory, one has to resort to nonperturbative means in order to address the critical behavior. Pruisken and co-workers were thus led to the dilute instanton gas approximation. 7 However, this approximation can only be justified in the weak-coupling limit, xx /(e 2 /h)ӷ1, and becomes uncontrolled in the critical region xx ϳe 2 /h. For this reason, no quantitative predictions for critical properties have been made within this approach. Another line of efforts was based on a mapping of the low-energy sector of Pruisken's model onto an antiferromagnetic superspin chain. 8 The superspin chain was also obtained by starting from the Chalker-Coddington network model of the IQHE. 9 However, attempts to find an analytical solution of the superspin chain problem remained unsuccessful.Recently, two papers appeared which may signify a breakthrough in the quest for the conformal critical theory of the IQHE. Zirnbauer 10 and, a few months later, Bhaseen et al.,11 proposed that this is a -model with the Wess-Zumino-Novikov-Witten ͑WZNW͒ term ⌫,where g belongs to a certain sy...

Wavefunction statistics and multifractality at the spin quantum Hall transition

“…The theory implies an exactly parabolic form of the multifractality spectrum. This was confirmed by a thorough numerical study of the wave function statistics at the QH transition [20].…”

“…In analogy with Anderson and quantum Hall transitions studied earlier [34,20,[35][36][37], we expect the distribution P(P q ) to become scale-invariant in the large-L limit. Figure 5 demonstrates that this is indeed the case.…”

“…Using (14), we classify the contributions to (20) according to the numbers of returns k 1 , k 2 to the edge f for the corresponding two paths. We want to show that only the contributions with k 1 + k 2 = 0, 2 are to be taken into account.…”

The instanton vacuum of generalized CPN−1 models

“…It is clear, however, that in order for the one-instanton approach to be successful the parameter must be a small quantity. By the same token, the expressions [[Á Á Á]] should be well approximated by inserting the leading order corrections as obtained from ordinary perturbative expansions [45] Here, the constant c has not yet been computed explicitly. The critical fixed point can then be obtained formally as an expansion in powers of ,…”

Interaction effects on two-dimensional fermions with random hopping

“…While some features of the physics of the sublattice model are certainly tied to the special chiral symmetry, others, particularly the multifractal nature of the extended wavefunctions 11,20,21,22,23,24,25 are believed to be more general features of (Anderson) localization critical points (see e.g. Refs.…”