It’s been a while! Part of the reason I’ve not written anything recently is that I’ve been busy preparing a grant proposal which has to be submitted in a few days. This means I’m begging the Swedish funding agency to give me money to spend on researching a new idea that I have been working on for a while. As part of this proposal, I am required to write a description of what I want to do that is understandable by people outside of physics, so I thought I’d share an edited version of it here. Maybe it’s interesting to read about something that might happen in the future, rather than things that are already well known. And it’s an idea that I’m pretty excited about because there’s some chance it might make a difference!
Computing technology is continuously getting smaller and more powerful. There is a rule-of-thumb, called Moore’s law, which encodes this by predicting that the computing power of consumer electronics will double every two years. So far, this prediction has been followed since microchips were invented in the 1970s. However, fundamental limits are about to be reached which will halt this progress. In particular, the individual transistors which make up the chips are becoming so small that quantum mechanical effects will soon start to dominate their operation and fundamentally change how they work. Removing the heat generated by their operation is also becoming hugely challenging.
A transistor is essentially just a switch that can be either on or off. At the present time, the difference between the on and off state is given by whether an electric current is flowing through the switch or not. If quantum mechanical effects start to dominate transistor operation, then the distinction between the on and off state becomes blurred because current flow becomes a more random process.
In this project, I will investigate a new method of making transistors, using the quantum mechanical properties of the electrons. The theoretical idea is to make two one-dimensional layers (for example, two nanowires) placed close enough to each other that the electrons in the material can interact with each other through Coulomb repulsion. If one of these nanowires has just a few electrons in it, while the other is almost full of electrons, then the electrons in the nearly empty wire can be attracted to the ‘holes’ in the nearly full wire, and they can pair up into new bound particles called excitons. What is special about these excitons is that they can form a superfluid which can be controlled electronically.
This can be made into a transistor in the following way. When the superfluid is absent, the two layers are quite well (although not perfectly) insulated from each other, so it is difficult for a current to flow between them. However, when the superfluid forms, one of the quantum mechanical implications is that it becomes possible to drive a substantial inter-layer current. This difference defines the on and off states of the transistor.
There are some mathematical reasons why one might expect that this cannot work for one-dimensional layers, but I have already demonstrated that there is a way around this. If the electrons can hop from one layer to the other, then the theorem which says that the superfluid cannot form in one dimension is not valid. What I will do next is a systematic investigation of lots of different types on one-dimensional materials to determine which is the best situation for experimentalists to look in for this superfluid. I will use approximate theories for the behaviour of electrons in nanowires or nanoribbons, carbon nanotubes, and core-shell nanowires to determine the temperature at which the superfluid can form for these different materials. When the superfluid is established, it can be described by a hydrodynamic theory which treats the superfluid as a large-scale object that can be described by simple equations that govern the flow of liquids. Analysing this theory will reveal information about the properties of the superfluid and allow optimisation of the operation of the switch. Finally, in reality, no material can be fabricated with perfect precision, so I will examine how imperfections will be detrimental to the formation of the superfluid to establish how accurate the production techniques need to be.
Another benefit of this superfluid is that it can conduct heat very efficiently. This means that it may have applications in cooling and refrigeration. I will also investigate the quantitative advantages that this may have over traditional thermoelectric materials. In both of these applications, the fact that the superfluid can exist in a one-dimensional material is a very advantageous factor for designing devices. In particular, because they are so small in two directions, it gives a huge amount of freedom for placing transistors or heat-conducting channels in optimal arrangements that would be impossible with two- or three-dimensional materials.
One final thing for some context: The picture at the top of the page shows a core-shell nanowire that was grown by some physicists in Lund, Sweden. It’s made out of two different types of semiconductor: Gallium antimonide (GaSb) in the core, and indium arsenic antimonide (InAsSb) in the shell. The core region is the nearly full layer that contains the ‘holes’, while the shell is the nearly empty layer with the electrons. The vertical white line on the left of the image is a scale bar that is 100nm long (that’s one ten-thousandth of a millimeter!) which shows that these wires are pretty small! (Picture credit: Ganjipour et al, Applied Physics Letters 101, 103501 (2012)).