Electric fields offer an innovative means of controlling condensed-matter systems. The approach has been applied to nanoscale oxide interfaces, for studying the physics of two-dimensional superconductors.
In recent years, the ability to engineer the interface between two oxides on the nanoscale with atomic-layer precision has led to remarkable advances in our knowledge of the electronic phenomena occurring at these interfaces. For instance, the transfer of charge at the interface between two ordinary oxide insulators, LaAlO3 and SrTiO3 (ref. 1), gives rise to a conducting two-dimensional gas of electrons that, at low temperatures, is a superconductor (it conducts electricity without resistance)2. Now Caviglia et al.3 (page 624 of this issue) show that an electric field can be used to modify the electron concentration of a transparent (clear) electron gas that forms at the LaAlO3/SrTiO3 interface (Fig. 1), allowing the emergence of two-dimensional superconductivity to be examined in unprecedented detail.
A major challenge in the study of these interfaces is identification of the source of the charge carriers (the electrons). One possible source is a 'polar catastrophe', in which the change in net layer charge (positive or negative) between the last TiO2 layer of the SrTiO3 structure and the first LaO layer of the LaAlO3 structure, drives an electronic reconfiguration that introduces charge carriers into the system1. Alternatively, they might be produced in a more mundane fashion by means of oxygen defects or the interdiffusion of cations between the LaAlO3 and SrTiO3 layers4. Caviglia and colleagues' work does not address this issue definitively, but it is not germane to the fundamental physics that they elucidate here.
Caviglia et al.3 use an electric field to reversibly pump charge carriers into or out of the interface, a process called charge doping. Impressively, this allows the authors to map out a region of the phase diagram of the interface state — a plot of temperature versus charge-carrier concentration that shows the phase boundary between the insulating state and the superconducting state. Because electric-field doping introduces no chemical or structural disorder at the interface, unlike chemical doping (in which the chemical composition is altered to introduce charges into the system), the authors are able to study the charge physics of a 'clean' system, without the obscuring, undesirable effects of disorder.
Interestingly, Caviglia et al. find no evidence of magnetism at the LaAlO3/SrTiO3 interface over the entire range of carrier concentrations studied, including densities of about 1013 electrons per cubic centimetre, for which a magnetic state was previously identified5. What they find instead is a superconducting, dome-shaped region in the phase diagram (see Fig. 3 on page 625), occurring over a range of charge-doping levels and temperatures that is consistent with the range over which doped SrTiO3 is known to be superconducting.
Caviglia and colleagues' pivotal finding is that this clean, disorder-free doping approach allows them to identify the quantum critical point that separates the superconducting state from the insulating state as a function of charge density. A detailed analysis of this critical region suggests quantum phase fluctuations that are reminiscent of insulator–superconductor quantum phase transitions, tuned by electric fields, seen in other two-dimensional systems6.
But where do we go from here? One clear role for the authors' approach of using an electric field to control superconductivity will be in elucidating the nature of phase transitions in other systems, particularly in systems in which the symmetry is broken by the presence of the interface. What's more, this approach will facilitate the development of nanoscale superconducting circuits, including the creation of rewritable arrays of Josephson junctions (two superconductors separated by a thin, non-superconducting region).
The authors' method will find even more wide-ranging applications. First, the electric-field control of interface properties, combined with the nearly unlimited possible combinations of oxides displaying diverse electronic properties, will allow other physical phenomena to be probed, including magnetism, orbital ordering and charge ordering7. The present interest in this approach is also partly due to its potential for achieving charge-mediated magnetoelectric coupling in composite multiferroic systems. These materials allow magnetism to be controlled in the solid state using electric fields. Furthermore, in its search for an eventual replacement for the transistor, the semiconductor industry is looking at exotic phenomena in various solid-state systems, including the control and manipulation of exotic ground states in novel, artificially structured materials. Electric-field control of these ground states could form the basic logic states ('0' and '1') of future information-processing technologies.
But most notably, Caviglia and colleagues' work3 marks the dawn of a new era in the design of superconductors consisting of materials created with atomic-layer precision. In these systems, huge strains can be imparted on the constituent monolayers assembled on top of one another. This could enable electronic-reconfiguration and orbital-ordering effects to enhance the superconductivity, which could be turned on and off simply by using an electric field to modulate the charge-carrier density. Engineered oxide films, in particular, offer a new means of searching for superconductivity that traditional methods lacked. Several decades ago, it was proposed that interfaces could be used to boost the temperature at which superconductivity can exist8,9, but it was also noted that this would be a challenge to achieve experimentally9. Studies such as that of Caviglia et al., together with advances in theory10, now show that this idea is becoming a reality.