Stabilization of Chemical-Vapor-Deposition-Grown WS₂ Monolayers at Elevated Temperature with Hexagonal Boron Nitride Encapsulation

Abstract

Chemical vapor deposition (CVD)-grown flakes of high-quality monolayers of WS2 can be stabilized at elevated temperatures by encapsulation with several layer hexagonal boron nitride (h-BN), but to different degrees in the presence of ambient air, flowing N2, and flowing forming gas (95% N2, 5% H2). The best passivation of WS2 at elevated temperature occurs for h-BN-covered samples with flowing N2 (after heating to 873 K), as judged by optical microscopy and photoluminescence (PL) intensity after a heating/cooling cycle. Stability is worse for uncovered samples, but best with flowing forming gas. PL from trions, in addition to that from excitons, is seen for covered WS2 only for forming gas, during cooling below ∼323 K; the trion has an estimated binding energy of ∼28 meV. It might occur because of doping level changes caused by charge defect generation by H2 molecules diffusing between the h-BN and the SiO2/Si substrate. The decomposition of uncovered WS2 flakes in air suggests a dissociation and chemisorption energy barrier of O2 on the WS2 surface of ∼1.6 eV. Fitting the high-temperature PL intensities in air gives a binding energy of a free exciton of ∼229 meV.

Improving the Optical Quality of MoSe2 and WS2 Monolayers with Complete h-BN Encapsulation by High-Temperature Annealing

“…By the calculation from Florian et al, a 5 Å interlayer gap between h-BN-MoSe 2 -h-BN can cause a trion binding energy difference of ∼15−20 meV. 49 In our earlier paper, 24 a CVD-grown WS 2 ML was seen to be more stable and had larger integrated PL at 300 K after a 300 K to 873 K to 300 K heating/cooling cycle with a top h-BN layer than without it and with nitrogen flow, rather than in ambient air or with flowing forming gas. However, the PL was still smaller than before the cycle, by ∼47%.…”

“…The peak PL intensity was larger after the annealing cycle than that before it, by ∼14 times for the exciton and ∼2.5 times for the trion. (The dark exciton effect lowers the overall PL intensity. − ) After the heating cycle, the peaks are red-shifted relative to those before, by 4.6 meV for the trion peak and 26.9 meV for the exciton. − These exciton peak red shifts correspond to an increase in the biaxial in-plane tensile strain of ∼0.04%–0.26% if it were all due to strain, , as noted above. Also, the trion peak red shift of ∼22 meV is more indicative of the reduction of trion binding energy.…”

“…45−47 ) After the heating cycle, the peaks are red-shifted relative to those before, by 4.6 meV for the trion peak and 26.9 meV for the exciton. 33−35 These exciton peak red shifts correspond to an increase in the biaxial in-plane tensile strain of ∼0.04%−0.26% if it were all due to strain, 24,48 as noted above. Also, the trion peak red shift of ∼22 meV is more indicative of the reduction of trion binding energy.…”

“…15 We have previously seen notable improvement in the stability and optical properties, notably PL, of chemical vapor deposition (CVD)grown WS 2 monolayers (MLs) after heating/cooling cycles, relative to the bare ML, by first fully covering the WS 2 ML with multilayer h-BN. 24 However, although PL was stronger with a top layer of h-BN than without it, the intensity still decreased in these samples. Markedly better improvements in both properties and stability might be expected with the better sealing of the edges of the TMD monolayer expected from full encapsulation of the ML by using top and bottom h-BN with larger areas than the TMD, prepared by employing established transfer methods.…”

“…Improved properties by thermal annealing and better stability at elevated temperatures might be expected by encapsulating the TMD ML using hexagonal boron nitride ( h -BN), which is known to be thermally stable at high temperature, before heating the TMD . We have previously seen notable improvement in the stability and optical properties, notably PL, of chemical vapor deposition (CVD)-grown WS 2 monolayers (MLs) after heating/cooling cycles, relative to the bare ML, by first fully covering the WS 2 ML with multilayer h -BN . However, although PL was stronger with a top layer of h -BN than without it, the intensity still decreased in these samples.…”