Glassy Dynamics of Electrons Near the Metal–Insulator Transition

Abstract

PrefaceThis review first describes the evidence that strongly suggests the existence of the metal-insulator transition (MIT) in a two-dimensional electron system in Si regardless of the amount of disorder. Extensive studies of the charge dynamics demonstrate that this transition is closely related to the glassy freezing of electrons as temperature T → 0. Similarities to the behavior of three-dimensional materials raise the intriguing possibility that such correlated dynamics might be a universal feature of the MIT regardless of the dimensionality. Contents 0.1 Introduction 1 0.2 Metal-insulator transition in two dimensions 3 0.3 Glassy freezing of electrons in two dimensions 15 0.4 Summary 28 0.5 Discussion 29 0.6 Acknowledgments 33 References 34 Free energy landscape Configuration space coordinate F Glass Fig. 0.1 The free-energy (F ) landscape of a glassy system. The horizontal axis represents the one-dimensional projection of the configurational coordinates of the degrees of freedom.Metal-insulator transition in two dimensions ¿ Lee and Ramakrishnan 1985). They have an additional advantage that all the relevant parameters, carrier concentration, disorder and interactions, can be varied relatively easily. This review will describe some of the work that has demonstrated that the 2DES in Si is an excellent model system not only for studying the MIT in two dimensions (2D), but also for investigating the nearly universal nonequilibrium behavior exhibited by a large class of both three-dimensional (3D) and 2D systems (e.g. spin glasses, supercooled liquids, granular films). In fact, the work on the 2DES in Si provides additional strong evidence that many such universal features are robust manifestations of glassiness, regardless of the dimensionality of the system. The dimensionality plays an important role in the MIT. In 2D, for example, the very existence of the metal and the MIT had been questioned for many years. Recently, considerable experimental evidence has become available in favor of such a transition. Some of that work has been described in several review papers (e.g. Kravchenko and Sarachik 2004) with a focus on very clean (low-disordered) samples. This review will first extend that discussion to the cases of higher disorder in order to (a) demonstrate that the 2D MIT is observed in all 2DES in Si regardless of the amount of disorder, (b) point out important similarities to the MIT in 3D systems, and (c) set the stage for the discussion of glassy dynamics near the MIT. As summarized below, the 2DES in Si exhibits all the main manifestations of glassiness: slow, correlated dynamics; nonexponential relaxations; diverging equilibration time, as T → 0; aging and memory. The results provide strong support to theoretical proposals describing the 2D MIT as the melting of a Coulomb glass (Dobrosavljević et al.

Low-Temperature 2D/2D Ohmic Contacts in WSe2 Field-Effect Transistors as a Platform for the 2D Metal–Insulator Transition

“…Furthermore, many novel materials behave effectively as two-dimensional (2D) systems, but the 2D MIT remains one of the most fundamental open problems in condensed matter science. [1][2][3][4][5][6][7] In fact, the 2D electron system in Si MOSFETs is the only system in which detailed and comprehensive investigations of this quantum (i.e. T = 0) phase transition have been carried out so far.…”

“…In general, however, the microscopic origin of the metal–insulator transition (MIT) in most materials of interest is not well understood. Furthermore, many novel materials behave effectively as two-dimensional (2D) systems, but the 2D MIT remains one of the most fundamental open problems in condensed matter science. − In fact, the 2D electron system in Si MOSFETs is the only system in which detailed and comprehensive investigations of this quantum (i.e., T = 0) phase transition have been carried out so far. On the basis of Si MOSFET studies, it has been suggested that there are three universality classes for the 2D MIT. , However, to test the universality of the observed behaviors, it is necessary to probe 2D systems in other types of materials and, in particular, beyond conventional semiconductor heterostructures.…”