No abstract

We present the results of field-and temperature-dependent resistivity and magnetization M measurements which show that colossal magnetoresistance in semimetallic EuB 6 is not associated with its bulk Curie temperature T C ϭ12.6 K, determined by a scaling analysis of M (H,T), but with a second phase transition at T M ϭ15.5 K. We propose that metallization occurs at T M via the overlap of magnetic polarons. The separation of the charge delocalization and bulk magnetic ordering transitions implies electronic and magnetic phase separation.The discovery of colossal magnetoresistance ͑CMR͒ in rare-earth manganites and manganese pyrochlores has renewed interest in the properties of magnetic polarons. 1 As has often been noted, CMR was also observed in a very different class of rare-earth magnets ͓EuO, 2 EuB 6 , 3-5 Gd 1Ϫx Eu x Se ͑Ref. 6͔͒ almost 30 years ago. It is generally believed that magnetic polarons are present in all of these materials, and that the sensitivity of polaronic transport to field is responsible for CMR. Although EuB 6 has historically been grouped with charge-doped rare-earth chalcogenides and oxides, in which the polaronic carriers are extrinsic, it has recently been shown that this intriguing material is a rare example of a low carrier density semimetal which also orders ferromagnetically. 7-9 We present here a detailed study of the field-and temperature-dependent resistivity and magnetization M of EuB 6 which suggests the presence of magnetic polarons in this system, and demonstrates that bulk magnetic order and metallization occur at different temperatures. This is a phenomenon unique to EuB 6 , and we will argue that it implies electronic and magnetic phase separation.EuB 6 crystallizes in the simple cubic CaB 6 structure. The electronic structure is derived primarily from the hybridization of the boron p bands, filled by electrons donated by the Eu ions, which are found to be divalent by both photoemission 10 and Mössbauer 11 measurements. Electronic structure calculations 7 indicate that small dilations of the boron octohedra cause overlap of the conduction and valence bands at the X points, rendering EuB 6 semimetallic. Shubnikov-de Haas and de Haas-van Alphen measurements 8,9 as well as angle-resolved photoemission ͑ARPES͒ measurements 10 confirm that EuB 6 is a semimetal. Quantum oscillation studies found that the sample studied here has an intrinsic carrier density of 1.2ϫ10 20 cm Ϫ3 . 9 Neutron-diffraction measurements on EuB 6 ͑Ref. 12͒ find that the spontaneous moment first appears near ϳ16 K, increases slowly as the temperature is lowered, and then displays quasi-mean-field behavior below ϳ12-13 K. The saturation moment of 6.9Ϯ0.2 B , measured at 1.5 K, is somewhat reduced from the nominal Eu 2ϩ value of 7.94 B , but is in good agreement with the 7.02 B predicted by bandstructure calculations. 7 Signatures of a second phase transition at 15.5 K have been found in the resistivity, the lowfield magnetization, and the zero-field specific heat. 13,14 Currently, there is no explanation fo...

The time-dependent fluctuations of conductivity σ have been studied in a two-dimensional electron system in low-mobility, small-size Si inversion layers. The noise power spectrum is ∼ 1/f α with α exhibiting a sharp jump at a certain electron density ns = ng. An enormous increase in the relative variance of σ is observed as ns is reduced below ng, reflecting a dramatic slowing down of the electron dynamics. This is attributed to the freezing of the electron glass. The data strongly suggest that glassy dynamics persists in the metallic phase.PACS Nos. 71.30.+h, 71.27.+a, 73.40.Qv The possibility of a metal-insulator transition (MIT) in two dimensions (2D) has been a subject of intensive research in recent years [1,2] but the physics behind this phenomenon is still not understood. It is well established that the MIT occurs in the regime where both Coulomb (electron-electron) interactions and disorder are strong. Theoretically, it is well known [3] that, in the strongly localized limit, the competition between electron-electron interactions and disorder leads to glassy dynamics (electron or Coulomb glass). Some glassy properties, such as slow relaxation phenomena, have been indeed observed in various insulating thin films [4][5][6]. Furthermore, recent work [7] has suggested that the critical behavior near the 2D MIT may be dominated by the physics of the insulator, leading to the proposals that the 2D MIT can be described alternatively as the melting of the Wigner glass [8], or the melting of the electron glass [9]. It is clear that understanding the nature of the insulator represents a major open issue in this field. Here we report the first detailed study of glassy behavior in a 2D system in semiconductor heterostructures. The glass transition is manifested by a very abrupt onset of a specific type of random-looking slow dynamics, together with other signs of cooperativity. Our results strongly suggest that the glass transition occurs in the metallic phase as a precursor to the MIT, in agreement with recent theory [10].While glassy systems exhibit a variety of phenomena [11], studies of metallic spin glasses have demonstrated [12] that mesoscopic, i. e. transport noise measurements are required in order to provide definitive information on the details of glassy ordering and dynamics. Fluctuations of conductivity σ as a function of chemical potential (or gate voltage V g , which controls the carrier density n s ) have been investigated extensively in the insulating regime [13] and near the MIT [14] in a 2D electron system in mesoscopic Si metal-oxide-semiconductor field-effect transistors (MOSFETs). The latter study indicated that Coulomb interactions dominate the physics near the MIT. In order to get reasonably reproducible fluctuations as a function of V g , it was necessary to make very slow sweeps of many hours over a very narrow range of V g . Thus, all measurements represented a time average. As a matter of fact, it had been already known [15] that, at fixed V g (or n s ), σ fluctuates as a function of t...

The critical conductivity exponent of Si:P changes from nearin zero field to 0.86+0. 15 in a magnetic field of 8 T, consistent with the theoretical expectation of 1. According to recent theory, similar behavior found earlier in Si:B, where spin-orbit scattering is strong, corresponds to the universality class for magnetic impurities. These measurements in Si:P thus constitute a clear determination of the critical conductivity exponent near the metal-insulator transition in the universality class for high magnetic field.Based largely on the elegant stress-tuning experiments of Paalanen et al. , ' the metal-insulator transition that takes place in doped semiconductors as a function of dopant concentration is thought to be a continuous, second-order phase transition.The critical behavior near the transition is governed by the dominant process, be it spin-orbit scattering, magnetic impurities, or magnetic field, that determines the symmetry of the system and its universality class. The critical exponent p, which characterizes the approach to the transition of the zerotemperature conductivity cr = oo[(n /. n, ) -I j", is generally assumed equal to the critical exponent v for the correlation length, an assumption that is strictly valid only in the critical region. Although most theoretical work yields a critical exponent of 1 under most circumstances, that result is not well established, particularly in the generic class where the spin-Aip and spin-orbit scattering rates and the Zeeman splitting are all small compared to the temperature. Based on work of Finkelshtein, various theoretical studies have indicated that the exponent should be close to I in the magnetic-field (MF) universality class and for the cases of magnetic impurities (MI) and spin-orbit (SO) scattering.Experiments have shown that p is generally close to 1 in the absence of a magnetic field in doped semiconductors as well as in the amorphous metal-semiconductor mixtures.Since the exponent is expected to also be 1 in the MF universality class, the effect of applying an external magnetic field cannot be tested in such systems.Thus, the finding of Ootuka, Matsuoka, and Kobayashi that the exponent changes in Ge:Sb from 0.9 at B =0 to 1 in B =4 T is suggestive but not conclusive. On the other hand, it has been known for some years that the critical conductivity exponent' p is between -, ' and -', in the silicon-based n-type doped semiconductors Si:P, "Si:As, ' ' Si:B, Si:P,As, ' and probably Si:Sb, ' as well as amorphous Cxa:Ar. ' It is only in these materials that one can determine whether or not the exponent changes to the expected value of 1 in a magnetic field. Indeed, magnetic tuning measurements by Shafarman et al. ' indicate a possible change in the critical conductivity exponent of Si:As in a magnetic field.In an earlier publication' we reported that the critical exponent in Si:B, where spin-orbit effects are strong, changes from 0.65 in zero field to approximately 1 in a magnetic field of 7.5 T. We attributed these results to the magnetic-field univ...