Photovoltaic Effect of WS2 Nanotube Devices

The bulk photovoltaic effect (BPVE) in WS2 nanotube devices is quite stable in terms of quality and quantity. The large decrease in the short-circuit current (Isc) with decreasing temperature cannot be explained simply by a decrease in the absorption coefficient, because the band gap is blue-shifted with decreasing temperature. Light with a wavelength of 632.8 nm (1.96 eV) almost resonates with the A-exciton of WS2 (a specific bonded state of an electron and a hole) and therefore produces the strongest signal.

The external quantum efficiency is as high as 1.3% in the power regime of 10 W.cm^-2 as inferred from the slope of a linear fit to the data. At a wavelength of 730 nm (1.70 eV), there is still a small photovoltaic response. Now, photons can only excite carriers through the indirect gap (1.45 eV, or wavelength 855 nm), and about 60% of the solar electromagnetic spectrum (photon energy above 1.45 eV) can be exploited to convert incident energy into electrical energy.

For the sake of completeness, such a dependence would not be expected for the Schottky barrier photovoltaic effect at the interface between WS2 nanotubes and metal contacts. This effect is maintained linearly, even up to 5 × 103 W.cm-2.

In 2D TMDs, the oblique incidence may produce photoelectric effects and photon drag effects. However, although these are also present in WS2 nanotubes, they play only a small role. In addition, if the production of light carriers is inhomogeneous, a net Dember effect may occur. However, despite some inhomogeneities (extended data), the results obtained with WS2 nanotubes are not qualitatively and quantitatively consistent with this mechanism. If inhomogeneities are accompanied by strain variations, the so-called flexible photoelectric effect may appear and induce BPVE.

However, the crossover from linear dependence to square root dependence points to a different source of BPVE. Among the many physical mechanisms that have been proposed, the "shift current" model may be applicable to WS2 nanotubes, as it does predict that the power-dependent crossover is a general feature caused by carrier-excited saturation. The shifting current arises from the Berry connection of the Bloch function when there is non-zero electron spontaneous polarization. Given that electron spontaneous polarization is non-zero in anisotropic nanotubes, the shift current may indeed be active in WS2 nanotube devices and could explain the observed BPVE.

A more specific, qualitative model of the shift current has been proposed for tubular structures made of heteropolar materials; in this model, the shift current is generated by carriers occupying the K and K′ valleys, quantified by momentum around the circumference of the tube. Because these valleys can be selectively resolved by using circularly polarized light, the shift current should change as it reverses the direction of circular polarization.

The symmetry reduction caused by the curved nature of nanotubes is a key factor. However, the shifted current model only considers the consequences of its Hamiltonian neutral momentum quantization, which is inadequate because curvature can induce many other effects, including effective magnetic fields, interband spin-lattice coupling, and changes in the orbital composition of Bloch bands.

Therefore, further theoretical considerations, including other curvature-related factors, are needed to clarify the specific origin of BPVE in WS2 nanotubes. Finally, researchers compared the intensity of BPVE in nanotubes with that of BPVE in ferroelectric materials, which also exhibit BPVE. To calculate j_sc in WS2 nanotube devices, the researchers assumed a solid cylindrical cross-section rather than a hollow tube.

Cited paper: Zhang Y J, Ideue T, Onga M, et al. Enhanced intrinsic photovoltaic effect in tungsten disulfide nanotubes[J]. Nature, 2019, 570(7761): 349-353.

• < Prev
• Next >