Non-conservative dynamics: a new driver in molecular electronics

Electrical current flow interacts with the atoms in a conductor and causes a variety of effects. The most familiar is Joule heating: this is how a lightbulb and a toaster works. Another example is electromigration: current exerts forces on atoms, much like a river pushes rocks in its way. These current-induced forces can make atoms migrate, leading to the formation of defects in the conductor that can ultimately cause it to break down.In recent years experimentalists have been able to produce conductors of truly atomic dimensions - the smallest possible in nature - such as an atomic chain or a molecule between two electrodes. The excitement of these novel structures is the vision of molecular-scale electronics: electronic devices and circuit elements down on the size scale of individual atoms and molecules.But the current densities in these tiny wires can be very large - up to ten to the power of fifteen amps per square metre - many orders of magnitude larger than in an ordinary lightbulb for example. Under these huge current densities everything is big: both Joule heating and current-induced forces in nanowires can be very large, and can blow the conductor to pieces. We have worked for years on the theory and modelling of these effects in atomic-scale devices, and recently we have made a discovery: interatomic bonding forces in atomic wires under current are non-conservative, meaning that they can do net work on the atoms when they are taken on a closed path. We demonstrated the consequences for the simplest possible geometry: an atomic chain with a bend. There, the current drives and accelerates the bend atom in an expanding orbit, creating an elemental, single-atom waterwheel.This discovery found an immediate resonance within the science community through two News-and-Views articles in the Nature journals, and immediately opens up two new directions for research. (i) The non-conservative forces are a new mechanism for energy transfer from the current to the atoms, quite distinct from Joule heating. We have already given reasons, in our original Nature Nanotechnology paper, to believe that this mechanism can be more powerful than Joule heating. Therefore, it could be the new non-conservative effects, and not ordinary heating, that are the key factor limiting the stability and functionality of molecular electronic devices. (ii) Our atomic waterwheel shows that these forces can also be used constructively to drive atomic-scale engines.Our present project will develop new simulation tools to investigate these two possibilities from first principles and model the effects of the non-conservative forces in nanoscale conductors. We will model and understand how much effective heating these forces can produce, and whether they have the ability to destroy atomic wires. We will explore another novel idea: that the non-conservative waterwheel effect, and not Joule heating, could serve as the activation mechanism for electromigration of atoms on surfaces and in interfaces. Finally, together with experimentalists in Leiden, we will investigate a possible device that can act as a current-driven atomic-scale motor: a molecule on a current-carrying surface with a freely rotating side group, with the surface current driving the rotor like a watermill. In addition to the experimental group in Leiden we have joined forces also with a leading theory group in Denmark, who have taken up our discovery and have started their own long-term programme of research into it. Although the goal is shared, our theoretical approaches are mutually complementary and, together with our experimental friends, we aim not only to explore these new phenomena, but also to create a new direction of research in our dynamic field.

Who will benefit? Through meeting the demands of the 2009 ITRS (see Impact plan for details) we will feed into the needs of the semiconductor industry. In year 4 of the project we will host a dissemination and exploitation workshop involving our immediate collaborators and invitees from Intel, Seagate and other industries, as well as delegates from Invest Northern Ireland. The workshop will be open to other interested parties, in particular researchers from transport, time-dependent density-functional theory and the atomistic simulation of non-adiabatic electron-ion dynamics. This meeting will strengthen the interdisciplinarity of the project by bringing these areas together. Our training plans for the PDRA and project student will make them competitive in pursuit of employment not only in academia but also in the computing industry and HPC. In the past, two of our EPSRC-supported researchers have found employment in Accelrys and ICHEC, Dublin. We anticipate that our simulations of the waterwheel and surface hopping atoms will appeal to the imagination of the public and hence there is the clear opportunity for public understanding and engagement. This can be communicated through the Communications Office of QUB, whose remit is to publicise the exciting new discoveries of its researchers. The School of Mathematics and Physics has a programme of lectures to secondary schools and we will use this as a further communication channel to the public and as a way of attracting young people into Mathematics and Physics. How will they benefit? The UK has the potential to take the lead in the exploitation of current-induced forces for the design of nanomotors. Specifically, the action of non-conservative forces has been opened up through our recent research. It is essential to maintain this momentum in order to keep the UK at the forefront and to maximise the opportunities for wealth creation. The quality of life is expected to benefit from the creative design of new nanotechnologies. We encourage our students and PDRAs to take early ownership of their projects and to lead the authoring of publications. Professional skills gained during the project arise from our in-house expertise in atomistic simulation. These will lead to broader applications in fluid dynamics, time-dependent Liouville dynamics and parallel high performance computing. We expect a significant use of graphical visualisation to benefit our student and PDRA as they bring the results before the public. What will be done? The first line of dissemination will be scientific papers, conferences and meetings with collaborators. As mentioned above, we also are planning a workshop aimed at dissemination and exploitation. We have planned a regular programme of visits to our collaborators with whom we are also planning to exchange students, as stated in the letter of support from Prof. Brandbyge. As is stressed in the letter of support from Prof. van Ruitenbeek, the broader interest in electromigration at Leiden will ensure that this theme in the project reaches the relevant community directly. The applicants will seek to identify those parts of the research which may have commercial potential, especially those dealing with molecular-scale motors. The Knowledge and Exploitation Unit at Queen's University will provide advice on protecting intellectual property. Between them, the investigators have experience in presentation and conference authoring, conference organisation and the dissemination of codes through appropriate channels.