UC Berkeley

Since its inception, the field of valleytronics has always carried the hopes of generating an additional degree of freedom to electronic and spintronic devices, allowing complementary usage of these functions in a single device and hence greatly increasing the processing power. However, several candidates that presumably possess potential in valley manipulation have since fall short of being practical.

Valleytronics is the control of population at “valleys” of certain qualifying materials, and valleys are essentially different conduction band minima or, in general, band gaps that can be distinctively selected by certain selection rules. In diamond-structured materials, this means forcing electrons out of one of the three pairs of degenerate conduction band minima along <100> by magnetic or electric field. For transition metal dichalcogenides, this means using circularly polarized light to selectively excite one of the two band gaps at the ±K points. Both systems require some sort of bias, low temperature, and/ or stringent material quality/ size, making them impractical. Following the hype on layered materials, from graphene to transition metal dichalcogenides, and anisotropic materials, such as black phosphorus, we decided to explore a close analogue in IV-VI systems. Anisotropy is rather common, but IV-VI systems have the unique additional attribute of having two local band gaps that are along high symmetry lines in reciprocal space. Because the global band gap is not at a high symmetry point, the two relevant band gaps becomes significant and the behaviour near these band edges are then well-defined and well-separated by symmetry due to the anisotropic nature of the bandstructure. Using bulk Tin (II) Sulfide (SnS) as a novel system, we found that the two band gaps along ΓX (smaller energy) and ΓY (larger energy) are respectively only excited by light polarized along the x (zigzag) and y (armchair) directions. The implication of this is that we now have a valleytronic system that is inherently polarized and can be activated via light-matter interaction without material or experimental conditional constraints. The non-degeneracy in the band gap values also means that we can decouple the valley degree of freedom with the excitation energy, making the system more robust.

2

Theoretical efforts have allowed us to understand this novel form of valleytronics behaviour. We found that the conduction band minima are predominantly made up of 5p orbitals from Sn, while the valence band maxima are mainly composed of 3p orbitals of S atoms. The exact symmetries of the orbitals that make up the bands allow the simplification of transition dipole moment integral analyses, which gives the selection rules. As a proof of concept that such valleytronics can be generalized, we also investigated the effect on SnSxSe1-x and showed both the tunability of both band gaps as well as the retainment of the valleytronics effect across the compositions. The band gap tunability is continuous, a consequence of the hybridization of Se 4p orbitals and S 3p orbitals. Similarly, the retainment of symmetry across all compositions means that the polarization degree, an important figure of merit in valleytronic system, is not compromised.

Finally, we fabricated optoelectronic devices out of SnS with the general configuration of photodetectors, where two pairs of electrodes were placed along the x and y axes respectively. We then subsequently probed the photocurrent of the devices with a combination of excitation wavelengths (above, in between, and below the two band gaps), excitation polarization (along x and y), and current direction (along x and y), obtaining distinctly different photocurrent response. This is the first demonstration of a device that utilizes the valleytronics effect to control current flow, paving the way to a multi-digit switch that can provide additional degrees of freedom per device. Through this multi-digit device, we also identified the key limitations and fundamental mechanisms contributing to the lack of perfect valleytronics selectivity in the photoresponse and propose critical areas of improvements of the system.

In conclusion, we identified a novel form of valleytronics that is more practical than its previous counterpart. Our demonstration includes light-matter interaction and photoresponse of SnS, and also generalization in the SnSxSe1-x system. We acknowledge the limitations of our system but remain optimistic about the potential of such a form of symmetry-driven valleytronics, specifically when key limitations such as growth methods are solved, and/ or the quest of multifunctional materials have been fulfilled.