The two-dimensional electron systems (2DES) in semiconductor heterostructures has been a hot bed for new physical phenomena, manifesting strong correlation physics, and a testing ground for novel theoretical ideas. Physics research on the system has in the past mostly focused on the correlation physics of ideally clean systems and the experiments have been designed and carried out on 2DES with the highest mobility attainable at the time. This research will focus on three new material/device structures to access previously inaccessible physical regimes of dilute 2DES with controllably tunably disorder. DC transport and microwave spectroscopy measurements will be made at low T down to the nuclear demagnetization refrigerator temperature range to search for the expected new phases of electron matter intermediate to the Wigner crystal in clean 2DES and the Anderson insulator in highly disordered 2DES. It will educate students in condensed-matter physics and the science of semiconductor materials. They will be equipped with hands-on expertize in experimental low temperature physics techniques and semiconductor device processing technology.
In modern day electronics, almost all device functions are performed by electrons that are confined to move in thin films, or along the interfaces, of semiconductors. Such so-called two-dimensional electrons in semiconductors, though behaving at normal ambient as an ordinary gas of charged particles, have been under properly designed experimental conditions a hot bed for new physical phenomena, manifesting the quantum physics of strongly correlated electrons, and a testing ground for novel theoretical ideas. This research will investigate the influence of disorder, which is inevitable in all semiconductors, and explore a new physical regime, where the interplay between disorder and electron correlation is expected to give rise to new phases of the electron matter. The research is multidisciplinary in nature; it will educate students to be fluent in condensed-matter physics and the science of semiconductor materials, and equip them with hands-on expertize in experimental low temperature physics techniques and the processing technology of semiconductor devices.
The main focus of this project is on the two-dimensional electron system (2DES) in the Si quantum well of Si-Si1-xGex heterostructures. Its objective is to optimize the semiconductor thin-film structure, the material growth condition, and the device architecture and fabrication procedure to produce a Si 2DES in which the landscape of the random disorder potential is sufficiently smooth that trapping and localization are not observed for areal density of the 2D electrons down to 1010/cm2 range. Observation of free-electron-like transport in this low density regime is a clear indication that the 2DES is sufficiently uniform for practical use to manufacture Si quantum-dot quantum computers, and that a new regime of strong electron-electron interaction physics is at hand. The research entails modeling, design, and device fabrication, in addition to carrying out transport measurements in high magnetic fields and at low temperatures. The experiments, their results, and the new physics they uncover have been published in the refereed journal articles listed under Journal Publications. Here, a brief description of the outcome is given under three headings: 1. Si 2DES over a million mobility Design, fabrication, and characterization of Si quantum well structures were systematically carried out to materialize an undoped Si/Si1-xGex heterostructure in which a 2DES in the Si quantum well has electron mobility μ=1.6*106cm2/Vs at density n=1.5*1011/cm2. The 2DES which resides in the Si well of the undoped heterostructure is capacitively induced using an insulated-gate field-effect transistor device structure. The device makes it now possible to access by transport experiments the low density regime down to n~1*1010/cm2. The random disorder potential landscape in the device is so smooth that it has been practical to use it to fabricate Si quantum computers. 2. 2D metal-insulator transition The work focuses on two directions to investigate whether the observed metallic conduction in 2D indeed manifests a metal-to-insulator quantum phase transition of the 2DES and what role disorder and electron-electron interaction play. The experiments concentrate, on the one hand, on 2D systems with mobilities indicating much less disorder in them than in the high mobility Si-MOSFETs. On the other hand, different device structures and different heterostructures of various materials have been employed to realize the 2D charge carriers in the 2D systems. The nature of the disorder, and consequently the disorder potential landscape the 2D electron sees, can be tuned from long range puddles-and-islands inhomogeneities resulting from modulation doping to short range disorder from surface roughness at the interfaces, or from atom size alloy impurities to extended defects from crystal lattice mismatch. The experimental results are: (a) Whereas in Si-MOSFETs with mobility μ~0.4*105cm2/Vs the transition is observed at a characteristic density nc~1*1011/cm2, in Si/Si1-xGex quantum well structures with Si 2D electron mobility μ~2.3*105cm2/Vs the transition is at nc~0.3*1011/cm2. In more general terms, as the single particle scattering rate 1/ τ decreases, nc decreases; and the characteristic interaction parameter rsc (rs is the dimensionless electron-electron interaction parameter given by the average electron-electron separation to Bohr radius ratio) increases, showing that interaction plays an important role. (b) The metallic conduction is a finite temperature anomaly, not a manifestation of a 2D metal-to-insulator quantum phase transition. In fact, at sufficiently low temperatures when the mobility at a fixed 2D density reaches a characteristic value, the conduction becomes insulator-like, irrespective of 1/τ. This result indicates the ground state to be an insulator. (c) The conduction in the insulator-like state is not thermally activated; The 2D charge carriers are not exponentially localized. 3. Absence of the ferromagnetic phase in Si 2DES in the large rs limit In Si-MOSFETs, the in-plane field magnetoresistance at metallic conduction densities first increases parabolicly with increasing B, and then saturates when the 2DES is fully Zeeman spin-polarized. The saturation magnetic field Bs shows, for densities n>1*1011/cm2, a linear dependence on n and extrapolates to Bs→o at n~nc. This result has been taken as evidence for an emergent ferromagnetic phase of the 2DES at nc. The Si 2DES in our Si-Si1-xGex quantum well structure samples has higher mobility and smoother disorder potential landscape. As a result, the density range for metallic conduction is extended using these samples down to nc~0.1*1011/cm2. The Bs (n) data for these samples, for n≥0.6*1011/cm2 follow the same linear dependence and agree with those from previous experiments. For n≤0.6*1011/cm2, however, Bs(n) deviates notably from being linear and can be fitted with Bs=0 at n=0 to a power law with an exponent 1.3 for n=0.3*1011/cm2 to the highest sample density of n=1.6*1011/cm2. In fact, Bs remains finite of ~0.5T at n=0.1*1011/cm2. No evidence of ferromagnetic instability in Si 2DES is detected in its dilute limit with rs≈30.
