Batch production of 6-inch uniform monolayer molybdenum disulfide catalyzed by sodium in glass.

Monolayer transition metal dichalcogenides (TMDs) have become essential two-dimensional materials for their perspectives in engineering next-generation electronics. For related applications, the controlled growth of large-area uniform monolayer TMDs is crucial, while it remains challenging. Herein, we report the direct synthesis of 6-inch uniform monolayer molybdenum disulfide on the solid soda-lime glass, through a designed face-to-face metal-precursor supply route in a facile chemical vapor deposition process. We find that the highly uniform monolayer film, with the composite domains possessing an edge length larger than 400 µm, can be achieved within a quite short time of 8 min. This highly efficient growth is proven to be facilitated by sodium catalysts that are homogenously distributed in glass, according to our experimental facts and density functional theory calculations. This work provides insights into the batch production of highly uniform TMD films on the functional glass substrate with the advantages of low cost, easily transferrable, and compatible with direct applications.

Abnormal gas-liquid-solid phase transition behaviour of water observed with in situ environmental SEM.

Gas-liquid-solid phase transition behaviour of water is studied with environmental scanning electron microscopy for the first time. Abnormal phenomena are observed. At a fixed pressure of 450 Pa, with the temperature set to -7 °C, direct desublimation happens, and ice grows continuously along the substrate surface. At 550 Pa, although ice is the stable phase according to the phase diagram, metastable liquid droplets first nucleate and grow to ~100-200 µm sizes. Ice crystals nucleate within the large sized droplets, grow up and fill up the droplets. Later, the ice crystals grow continuously through desublimation. At 600 Pa, the metastable liquid grows quickly, with some ice nuclei floating in it, and the liquid-solid coexistence state exists for a long time. By lowering the vapour pressure and/or increasing the substrate temperature, ice sublimates into vapour phase, and especially, the remaining ice forms a porous structure due to preferential sublimation in the concave regions, which can be explained with surface tension effect. Interestingly, although it should be forbidden for ice to transform into liquid phase when the temperature is well below 0 °C, liquid like droplets form during the ice sublimation process, which is attributed to the surface tension effect and the quasiliquid layers.

Influence of water vapor on the electronic property of MoS2 field effect transistors.

The influence of water vapor on the electronic property of MoS2 field effect transistors (FETs) is studied through controlled experiments. We fabricate supported and suspended FETs on the same piece of MoS2 to figure out the role of SiO2 substrate on the water sensing property of MoS2. The two kinds of devices show similar response to water vapor and to different treatments, such as pumping in the vacuum, annealing at 500 K and current annealing, indicating the substrate does not play an important role in the MoS2 water sensor. Water adsorption is found to decrease the carrier mobility probably through introducing a scattering center on the surface of MoS2. The threshold voltage and subthreshold swing of the FETs do not change obviously after introducing water vapor, indicating there is no obvious doping and trap introducing effects. Long time pumping in a high vacuum and 500 K annealing show negligible effects on removing the water adsorption on the devices. Current annealing at high source-drain bias is found to be able to remove the water adsorption and set the FETs to their initial states. The mechanism is proposed to be through the hot carriers at high bias.

Remarkable influence of slack on the vibration of a single-walled carbon nanotube resonator.

We for the first time quantitatively investigate experimentally the remarkable influence of slack on the vibration of a single-walled carbon nanotube (SWCNT) resonator with a changeable channel length fabricated in situ inside a scanning electron microscope, compare the experimental results with the theoretical predictions calculated from the measured geometric and mechanical parameters of the same SWCNT, and find the following novel points. We demonstrate experimentally that as the slack s is increased from about zero to 1.8%, the detected vibration transforms from single-mode to multimode vibration, and the gate-tuning ability gradually attenuates for all the vibration modes. The quadratic tuning coefficient α decreases linearly with 1/√s when the gate voltage V(g)dc is small and the nanotube resonator operates in the beam regime. The linear tuning coefficient γ decreases linearly with 1/ (4√S) when V(g)dc has an intermediate value and the nanotube resonator operates in the catenary regime. The calculated α and γ fit the experimental values of the even in-plane mode reasonably well. As the slack is increased, the quality factor Q of the resonator linearly goes up, but the increase is far less steep than that predicted by the previous theoretical study. Our results are important to understand and design resonators based on CNT and other nanomaterials.

The intrinsic origin of hysteresis in MoS2 field effect transistors.

We investigate the hysteresis and gate voltage stress effect in MoS2 field effect transistors (FETs). We observe that both the suspended and the SiO2-supported FETs have large hysteresis in their transfer curves under vacuum which cannot be attributed to the traps at the interface between the MoS2 and the SiO2 or in the SiO2 substrate or the gas adsorption/desorption effect. Our findings indicate that the hysteresis we observe comes from the MoS2 itself, revealing an intrinsic origin of the hysteresis besides some extrinsic factors. The fact that the FETs based on thinner MoS2 have larger hysteresis than that with thicker MoS2 suggests that the surface of MoS2 plays a key role in the hysteresis. The gate voltage sweep range, sweep direction, sweep time and loading history all affect the hysteresis observed in the transfer curves.