Metal-oxide-semiconductor field-effect transistors (MOSFETs) on CdTe/HgTe/CdTe heterostructures are fabricated with silicon dioxide gate insulators. In these devices, the density of the quasi two-dimensional electron gas in the HgTe quantum well can be tuned in a wide range. In low magnetic fields we observe beating patterns in the Shubnikov-de Haas oscillations that render possible the determination of the coefficient α of the Rashba term in the Hamiltonian as a function of electron density. This coefficient consistently describes the splittings observed in cyclotron resonance in low magnetic fields.

No abstract

Ion milling has been used to convert molecular beam epitaxy vacancy-doped CdxHg1−xTe from p- to n-type. Electron beam induced current and remote electron beam induced current (REBIC) measurements have been performed to study the pn junction depth and lateral extension dependence on the milling time, milling current, and vacancy concentration. The conversion depth is linear with the milling time and current and inversely proportional to the vacancy concentration in layers thinner than 10 μm. This shows that filling of Hg vacancies in this region during conversion is limited by the rate of supply of extra Hg from the milling. The lateral extension also increases linearly with the milling time, the ratio of the lateral extension to the depth being ∼0.5. One can therefore use REBIC on the top surface to determine the junction depth, which greatly simplifies the measurement and does not destroy the diodes.

It is well known that the crystalline quality of CdxHgl_xTe grown by molecular beam epitaxy is critically dependent on the substrate temperature. The optimal growth temperature has been identified immediately below the crossing of the Te-rich phase boundary, that is just below the temperature range where Te precipitation occurs in the layer. It is potentially very useful to be able to predict the optimal temperature and its variation with other growth parameters, but no general guidelines for this can be found in the literature. We have studied experimentally the variation of the optimal growth temperature with Hg flux, Cd mole fraction and growth rate. These results are compared with a thermodynamical model published previously by Gailliard. We find that the modeled position of the phase boundary coincides well with the observed variations in optimal growth temperature for growth on Te-terminated surfaces, within the uncertainties of available thermodynamical constants. We show that the optimal substrate temperature depends mainly on the Hg flux and Cd mole fraction, while the dependence on growth rate can be neglected in practical molecular beam epitaxy conditions. The experimental observation of optimal layer quality at the phase boundary could suggest the existence of an adsorbed layer of Te, acting as a reservoir for Te atoms and reducing the supersaturation of the growth reaction. Simultaneous growth on the (211)B and (100) orientations reveals a clear, although not very large, difference in optimal growth temperature and Cd incorporation, indicating a difference in growth kinetics. This can be accounted for in the thermodynamical model by condensation and evaporation coefficients.

No abstract