UC Berkeley

In this thesis, I report my work on two different projects in the field of energy-efficient computation:

Part 1 (Device Physics): Tunnel Field-Effect Transistors (tFETs) are one of the candidate devices being studied as energy-efficient alternatives to the present-day MOSFETs. In these devices, the preferred switching mechanism is the alignment (ON) or misalignment (OFF) of two energy levels or band edges. Unfortunately, energy levels are never perfectly sharp. When a quantum dot interacts with a wire, its energy level is broadened. Its actual spectral shape controls the current/voltage response of such transistor switches, from on (aligned) to off (misaligned). The most common model of spectral line shape is the Lorentzian, which falls off as reciprocal energy offset squared. Unfortunately, this is too slow a turnoff, algebraically, to be useful as a transistor switch. Electronic switches generally demand an ON/OFF ratio of at least a million. Steep exponentially falling spectral tails would be needed for rapid off-state switching. This requires a new electronic feature, not previously recognized: narrowband, heavy-effective mass, quantum wire electrical contacts, to the tunneling quantum states.

Part 2 (Systems Physics): Optimization is built into the fundamentals of physics. For example, physics has the principle of least action, the principle of minimum power dissipation, also called minimum entropy generation, and the adiabatic principle, which, in its quantum form, is called quantum annealing. Machines built on these principles can solve the mathematical problem of optimization, even when constraints are included. Further, these machines become digital in the same sense that a flip–flop is digital when binary constraints are included. A wide variety of machines have had recent success at approximately optimizing the Ising magnetic energy. We demonstrate that almost all those machines perform optimization according to the principle of minimum power dissipation as put forth by Onsager. Moreover, we show that this optimization is equivalent to Lagrange multiplier optimization for constrained problems.