High-performance lateral MoS2-MoO3 heterojunction phototransistor enabled by in-situ chemical-oxidation

Abstract

Construction of in-plane p-n junction with clear interface by using homogenous materials is an important issue in two-dimensional transistors, which have great potential in the applications of next-generation integrated circuit and optoelectronic devices. Hence, a controlled and facile method to achieve p-n interface is desired. Molybdenum sulfide (MoS 2 ) has shown promising potential as an atomic-layer ntype semiconductor in electronics and optoelectronics. Here, we developed a facile and reliable approach to in-situ transform n-type MoS 2 into p-type MoO 3 to form lateral p-n junction via a KI/I 2 solution-based chemical oxidization process. The lateral MoS 2 /MoO 3 p-n junction exhibits a highly efficient photoresponse and ideal rectifying behavior, with a maximum external quantum efficiency of~650%, 3.6 mA W −1 at 0 V, and a light switching ratio of~10 2 . The importance of the built-in p-n junction with such a high performance is further confirmed by high-resolution photocurrent mapping. Due to the high photoresponse at low source-drain voltage (V DS ) and gate voltage (V G ), the formed MoS 2 /MoO 3 junction p-n diode shows potential applications in low-power operating photodevices and logic circuits. Our findings highlight the prospects of the local transformation of carrier type for high-performance MoS 2 -based electronics, optoelectronics and CMOS logic circuits.

Controlled Differentiation of Human Neural Progenitor Cells on Molybdenum Disulfide/Graphene Oxide Heterojunction Scaffolds by Photostimulation

“…MoO 3– x exhibit remarkably deep lying electronic states and are strongly n-doped by oxygen vacancies . The relative band position between MoS 2 and MoO 3– x domains can be estimated via two approaches: (1) by their electron affinity (the electron affinity of MoS 2 (χ MoS 2 ) and MoO 3 (χ MoO 3 ) were reported as ∼4.2 and ∼6.7 eV, respectively, while the electron affinity and work function of MoO 3 can be engineered based on oxygen vacancies , ) and (2) by their DTS spectra and Mott–Schottky plots, as reported by Huang et al In this study, the band gap energies of MoS 2 and MoO 3– x were calculated by DTS analysis (Figure ) which then estimated the approximate positions of their conduction band (CB) and valence band (VB) using the literature, as reported elsewhere . V0 is the built-in potential, and E (with the direction from n-MoO 3– x to p-MoS 2 ) is the built-in electric field. , Due to the Fermi level difference between n-MoO 3– x and p-MoS 2 and achieving equilibrium, the electrons (as majority charge carriers) diffuse from n-MoO 3– x to p-MoS 2 , and simultaneously, the holes in p-MoS 2 (as majority charge carriers) move to n-MoO 3– x , leading to the formation of a space-charge region (the depletion region) and a built-in electric field at the contact interface. , The built-in electric field, E , prevents carriers from diffusing and makes them drift in the opposite direction to diffusion, finally, when the drift current matches the diffusion current, the heterojunction comes to thermal equilibrium with a unified Fermi level. , Under LED excitation, the photoexcited electrons and holes in junctions are quickly separated and driven into n-MoO 3– x and p-MoS 2 , respectively, under the acceleration of E , which gives rise to the generation of the photoelectron current in the MoS 2 –MoO 3– x heterojunction. , In fact, the transfer of photoexcited electrons of MoS 2 to the CB of MoO 3– x , when LED irradiates the MoS 2 –MoO 3– x heterojunction implies that this interface charge transfer induces the separation of photoexcited e – –h + pairs and results in enhanced photoelectrical activity for the MoS 2 –MoO 3– x heterojunction. , If photoexcited electrons were transferred fast to hNPCs, they would be significantly differentiated into neurons (see Figure B), but this did not happen.…”

“…Before investigating the mechanism of lack of neural differentiation of hNPC on the MoS 2 –MoO 3– x heterojunction, it is necessary to mention a point here. MoO 3– x domains with rich oxygen vacancies are highly reactive, have strong antioxidant activity, and are prone to be oxidized by the electron transfer to free radicals (ROS), which the electron transfer to ROS (as the electron sink) leads to the formation of strong Mo–O bonds (2.395 eV), the induction of the high density of p-type doping, and the decrease of the surface charge density in ultrathin MoS 2 –MoO 3– x heterojunctions. ,,, Compared to darkness, under LED excitation, the antioxidant activity of MoO 3– x domains considerably enhanced due to LSPR absorption. ,, Therefore, under LED excitation, the photogenerated hole–electron pairs were formed on the surface of the MoO 3– x domains due to LSPR absorption; then, the photoexcited electrons on MoO 3– x domains are transferred to free radicals, and simultaneously, the photogenerated holes led to the oxidation of the MoO 3– x domains and the formation of more Mo–O bonds, which acted as the injection sites of p-type doping (Figure B). ,, Hence, in photostimulation, the more p-type doping injection led to a decrease in the photoelectron current and the prevention of the rapid and adequate transfer of photo-generated electrons to 1T-MoS 2 domains and then to hNPCs. ,, In photostimulation, the competition of the photo-generated electron adsorption was between hNPCs and free radicals, in which the lack of hNPC differentiation into neurons showed that free radicals by rapidly absorbing electrons increased the hole carriers (p-type dopings), resulting in glial differentiation. Also, the charge traps present in 2D ultrathin MoS 2 –MoO 3– x heterojunctions (due to their surface roughness and anisotropic physical properties), especially at the interfaces (p–n junctions), can act as the scattering centers and prevent the fast electron transfer to hNPCs, because the carrier mobility is influenced by carrier injection and scattering .…”

“…Despite the low neuronal differentiation of hNPCs on MoS 2 −MoO 3−x nanosheets, the neurons possessed more elongated cell shapes and neurite outgrowths (see Figure 9A-photostimulation). In photostimulation, the lack of favorable neuronal differentiation on the rGO substrate was due to the absence of a desirable light sensitivity of rGO respectively, while the electron affinity and work function of MoO 3 can be engineered based on oxygen vacancies 66,68 ) and…”

Dry Exfoliation of Large-Area 2D Monolayer and Heterostructure Arrays

“…For example, it has been reported that KI/I 2 solution could function as an oxidant to convert MoS 2 to MoO 3 . 26 Similarly, iodine complexes after KI/I 2 etching are reported to be the dopant absorbing on 2D materials' surface. 27,28 In addition, the involvement of solution etching also contradicts the primary motivation of using the dry tape exfoliation method.…”

“…Hence, solution-based etching process (typically KI/I 2 solution) is required to chemically remove the Au tapes, ,,− which leads to unavoidable contamination or damage to the delicate 2D lattice. For example, it has been reported that KI/I 2 solution could function as an oxidant to convert MoS 2 to MoO 3 . Similarly, iodine complexes after KI/I 2 etching are reported to be the dopant absorbing on 2D materials’ surface. , In addition, the involvement of solution etching also contradicts the primary motivation of using the dry tape exfoliation method.…”