Electrical conductivity of metallic Si:B near the metal-insulator transition

Abstract

The conductivity has been measured between 55 mK and 4.2 K in zero field and in magnetic fields up to 7.5 T of a series of uncompensated p-type Si:B samples with dopant concentrations near the critical concentration for the metal-insulator transition. Acceptor wave functions, which are derived in silicon from the degenerate lightand heavy-hole J =valence-band maxima at k =0 and a spin-orbit-split J = 2 band, are quite different from donor wave functions associated with the six degenerate conduction-band minima at different equivalent points in the Brillouin zone. Despite this, the conductivity of Si:8 is found to be quite similar in many ways to that of Si:P. The critical conductivity exponent for Si:8 is close to 2 as in Si:P and Si:As, rather than having the expected value of 1. The correction to the zero-temperature conductivity arising from electron-electron interactions is comparable in size, and the temperature dependence of the conductivity in various fixed magnetic fields is also found to be quite similar. For the range of dopant concentrations and experimental parameters of these investigations, the only important experimental difference between the two materials is the sign and size of the magnetoresistance. In contrast with Si:P, which has both positive and negative components, the magnetoresistance of Si:8 is positive for all temperatures and magnetic fields studied. We attribute this to the strong spin-orbit scattering in p-type silicon associated with the degenerate valence bands.

Insulator-to-metal transition in vanadium supersaturated silicon: variable-range hopping and Kondo effect signatures

“…This confirms that the underneath silicon substrate is not playing any role in the transport at low temperatures. Moreover, for temperatures below 25 K these samples present a temperatureindependent n s value, as expected for an impurified semiconductor in the metallic phase [29]. We thus conclude that layers implanted with 10 13 and 10 14 cm −2 doses are in the insulating phase, whereas samples implanted with 10 15 and 10 16 cm −2 are in the metallic phase.…”

Dynamical mean–field studies of metal–insulator transitions

“…In the presence of a strong magnetic field the conductivity vanishes with an exponent µ ≈ 1, in both compensated an uncompensated semiconductors (Dai et al 1991(Dai et al , 1992. The presence of spin-orbit (SO) coupling (Dai et al 1991(Dai et al , 1992 does not seem to be a determining factor for the behavior of the conductivity. Both Si:B (where SO coupling is strong) and Si:P (where SO coupling is weak) behave qualitatively in the same way.…”

Metal-insulator transition of isotopically enriched neutron-transmutation-doped70Ge:Gain magnetic fields

“…In their approach, the ratio of the resistance at 4.2 K to that at 300 K is used to determine the concentration. 17 The dashed curve in Fig. 2 is for B = 0, which merely expresses Eq.…”

“…Therefore, comparison of µ with and without the time-reversal symmetry, i.e., with and without external magnetic fields becomes important. Experimentally, µ ≈ 1 has been found for magnetic inductions B on the order of one tesla for nominally uncompensated semiconductors: Ge:Sb, 15,16 Si:B, 17 and Si:P (Ref. 18).…”

“…This result was obtained from precisely doped samples with a perfectly random distribution of impurities; our 70 Ge:Ga samples were prepared by neutrontransmutation doping (NTD), in which an ideally random distribution of dopants is inherently guaranteed down to the atomic level. [20][21][22][23] For the case of melt-(or metallurgically) doped samples that have been employed in most of the previous studies, [3][4][5][15][16][17][18][19] the spatial fluctuation of N due to dopant striations and segregation can easily be on the order of 1% or more across a typical sample for the four-point resistance measurement (length of ∼5 mm or larger), 24 and hence, it will not be meaningful to discuss physical properties in the critical regime (e.g., |N/N c − 1| < 0.01), unless one evaluates the macroscopic inhomogeneity in the samples and its influence on the results. A homogeneous distribution of impurities is important also for experiments in magnetic fields.…”