Anomalous parallel-field negative magnetoresistance in ultrathin films near the superconductor-insulator transition

Abstract

A parallel field negative magnetoresistance has been found in quench-condensed ultrathin films of amorphous bismuth in the immediate vicinity of the thickness-tuned superconductorinsulator transition. The effect appears to be a signature of quantum fluctuations of the order parameter associated with the quantum critical point.Quantum phase transitions (QPTs) are brought about by the variation of an external parameter of the Hamiltonian of a system, which changes the ground state [1]. The superconductor-insulator (SI) transition in two dimensions (2D), tuned by disorder or magnetic field, is believed to be a quantum phase transition. The understanding of the SI transition as a QPT has been inferred from the successful analysis of transport data using finite size scaling. However recent data on the field-and disorder-tuned SI transitions suggest the existence of a finite intermediate regime of metallic behavior not anticipated by the theory [2]. Subsequent explanations of this regime have included a metallic Bose glass or a Bose metal [3], metallicity produced by the influence of dissipation [2], and effects resulting from the influence of fermionic excitations not included in boson models [4]. In some instances the metallic regime may be attributable to the electrons not cooling. Because of these complications, it would be useful if there were an explicit signature of quantum fluctuations that could serve as an indicator of the SI transition. A recent calculation [5] appears to offer this possibility. Employing a perturbative approach, a negative correction to the parallel field magnetoresistance (MR) attributable to quantum fluctuations has been found near the parallel-field SI transition of films (and wires). The total negative MR results from the "Aslamazov-Larkin" correction being overwhelmed by negative contributions from the "density of states" and "Maki-Thompson" terms. In this letter we report an anomalous, parallel-field negative MR whose occurrence is correlated with the thickness-tuned SI transition of ultrathin homogeneous films. This effect may derive from corrections to the conductivity associated with quantum fluctuations even though the effect is found near the condition of critical disorder rather than critical parallel magnetic field.Resistance measurements were made using a bottom loading Kelvinox 400 dilution refrigerator, employing four-probe techniques. Electrical leads were filtered at room temperature using π-section filters with a cutoff frequency of about 10 Hz. Power dissipation in the measurement process was kept below 1 pW. The substrate was mounted on a sample holder that could be transferred between the mixing chamber of the refrigerator and an attached ultra-high vacuum growth chamber using a liquid-helium-cooled transfer stick [6]. In these experiments the plane of the substrate, mounted on a rotatable sample holder, was restricted to be close to the nominally parallel orientation to accommodate additional heat sinking needed to facilitate cooling below 0.1K.Films were grown on su...

Vortex-induced negative magnetoresistance and peak effect in narrow superconducting films

“…There are many works [35][36][37][38][39][40]57] where measured R(H) has shape similar to curve R a in Fig. 1.…”

Electrostatic and parallel-magnetic-field tuned two-dimensional superconductor-insulator transitions

“…The physical evidence for such a regime is that resistances become independent of temperature at the lowest temperatures both in perpendicular field tuned transitions 24 and in thickness-tuned transitions. 25 In the case of the magnetic field experiments, the metallic regime abruptly disappears as the magnetic field is reduced, leading to true superconductivity. This is further evidence of possible sample inhomogeneity.…”

Length dependence of the resistance in graphite: Influence of ballistic transport

“…Among those materials are the ones, which show ballistic transport, at least for a given device channel length like in some carbon nanotubes samples [1][2][3]. The direct measurement of ℓ is possible, for example, with a scanning microscope through the scaling of the channel resistance [4] or with a multi terminal method [5]. In Ref.…”