There are a tremendous number of conduction electrons in a metal. Although they have negative charges and feel repulsive forces from each other, such an effect of electron interaction does not significantly change the electronic states in most materials, because of the screening effect. The electronic band structure derived from one-body models can predict the electronic properties of a material.
However, in strongly correlated materials, electron interaction is not negligible because of some reason (heavy electron mass, low dimensionality, and so on), and exhibits exotic quantum many-body effects. Those materials can be insulating because of electron interaction (the Mott insulator) even if the band theory predicts that they are metal. The key parameters in such materials are the density and the kinetic energy of electrons (the latter corresponds to the strength of electron correlation). With varying those parameters, drastic metal-insulator transitions occur, and superconductivity often emerges âbetweenâ the insulating and metallic states. High-temperature superconducting cuprates are the typical case. The mechanism of the superconductivity has been intensively investigated for decades and is yet to be fully clarified.
A straightforward experiment to study such a system is to tune the density and the kinetic energy of electrons precisely and see what happens in a single sample. However, such a measurement is challenging. In experiments, density and kinetic energy are usually controlled with chemical doping and pressure, respectively. In most materials, we can sufficiently control only either of them. Besides, chemical doping introduces impurities and sample-to-sample dependence.
One approach to achieve such an experiment is to construct a highly controllable artificial system (which is called a quantum simulator) and reproduce the strongly correlated systems. This has been developed using ultracold atoms in optical lattices. We have recently attempted another method that we fabricate a transistor device based on a highly controllable organic strongly correlated material. As a result, electrostatic doping with a gate voltage and application of strain with bending the substrate can reversibly modify the density and kinetic energy of the correlated electrons, and switch on and off the superconductivity in a single sample (Fig. 1).
Fig. 1. Schematic of the manipulation of the organic strongly correlated transistor. A gate voltage varies the number (density) of electrons, and a mechanical strain modifies the distance between molecules, resulting in a change in the kinetic energy of electrons. Figure courtesy Yoshitaka Kawasugi.
We fabricated an electric-double-layer transistor using an organic Mott insulator Îș-(BEDT-TTF)2Cu [N(CN)2]Cl and flexible PET substrate. The half-filled carrier density of this material is several times smaller than those of typical inorganic materials. Therefore, a relatively weak gate voltage can widely vary the carrier concentration across zero doping (approximately ±20%). At the same time, the material is very sensitive to pressure due to its soft lattice. A small strain owing to the bending of the substrate can induce insulator-superconductor transitions). In this study, we combine the gating and bending controls in a single sample and investigated the conditions for the superconductivity in the strongly correlated material.
The device showed that all of the electron doping, hole doping, and bending strain could induce superconductivity, and the superconducting state surrounds the insulating state in the gate voltage-strain phase diagram (Fig. 2). The superconducting region in the phase diagram is quite asymmetric against doping (electron-hole asymmetry), implying that the band structure without electron correlation also affects the correlation-induced superconductivity. The observed phase diagram also showed many unexpected features such as an abrupt first-order superconducting transition under electron doping, a recurrent insulating phase in the heavily electron-doped region, and a nearly constant superconducting transition temperature in a wide parameter range.
Fig. 2. Gate voltage and strain dependence of resistivity in our organic strongly correlated transistor. The red and dark blue regions correspond to the insulating and superconducting states (h-SC and e-SC denote the hole-doped and electron-doped superconducting states, respectively). One sees that the phase diagram is doping-asymmetric, and the right-hand side appears anomalous. Figure courtesy Yoshitaka Kawasugi.
The results show not only the intrinsic and intriguing physics of the strongly correlated electrons but also the high controllability of the molecular-based organic strongly correlated device.
These findings are described in the article entitled Two-dimensional ground-state mapping of Mott-Hubbard system in a flexible field-effect device, recently published in the journal Science Advances.